Upload folder using huggingface_hub

README.md

@@ -0,0 +1,9 @@
+---
+library_name: peft
+---
+## Training procedure
+
+### Framework versions
+
+
+- PEFT 0.4.0

adapter_config.json

@@ -0,0 +1,26 @@
+{
+  "auto_mapping": null,
+  "base_model_name_or_path": "/model_zoo/Vivid-7B-base",
+  "bias": "none",
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "lora_alpha": 64,
+  "lora_dropout": 0.05,
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "r": 32,
+  "revision": null,
+  "target_modules": [
+    "gate_proj",
+    "k_proj",
+    "up_proj",
+    "q_proj",
+    "down_proj",
+    "v_proj",
+    "o_proj"
+  ],
+  "task_type": "CAUSAL_LM"
+}

adapter_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5b41c5203e7bea743d0843d653bf92094ae79363e8113b05a4cae5ff1bd97825
+size 167927754

checkpoint-300/README.md

@@ -0,0 +1,10 @@
+---
+library_name: peft
+---
+## Training procedure
+
+### Framework versions
+
+- PEFT 0.4.0
+
+- PEFT 0.4.0

checkpoint-300/adapter_config.json

@@ -0,0 +1,26 @@
+{
+  "auto_mapping": null,
+  "base_model_name_or_path": "/model_zoo/Vivid-7B-base",
+  "bias": "none",
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "lora_alpha": 64,
+  "lora_dropout": 0.05,
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "r": 32,
+  "revision": null,
+  "target_modules": [
+    "gate_proj",
+    "k_proj",
+    "up_proj",
+    "q_proj",
+    "down_proj",
+    "v_proj",
+    "o_proj"
+  ],
+  "task_type": "CAUSAL_LM"
+}

checkpoint-300/adapter_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c5248907b972eb9eef54d8796e7a990b4dcd83dc26fecbc9b8c6777b0ea1217b
+size 167927754

checkpoint-300/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:297260dc0688480fdc1b1d063f8b141060ef419e224266481e67de2679298da1
+size 167832688

checkpoint-300/config.json

@@ -0,0 +1,43 @@
+{
+  "_name_or_path": "/model_zoo/Vivid-7B-base",
+  "architectures": [
+    "MistralForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "bert_type": "raw_bert_layer:12",
+  "bos_token_id": 1,
+  "compress_type": "mean",
+  "eos_token_id": 2,
+  "freeze_mm_mlp_adapter": false,
+  "hidden_act": "silu",
+  "hidden_size": 4096,
+  "image_aspect_ratio": "pad",
+  "image_grid_pinpoints": null,
+  "image_processor": "./llamavid/processor/clip-patch14-224",
+  "initializer_range": 0.02,
+  "intermediate_size": 14336,
+  "max_position_embeddings": 32768,
+  "max_token": 4096,
+  "mm_hidden_size": 1024,
+  "mm_projector_type": "mlp2x_gelu",
+  "mm_use_im_patch_token": false,
+  "mm_use_im_start_end": false,
+  "mm_vision_select_feature": "patch",
+  "mm_vision_select_layer": -2,
+  "mm_vision_tower": "/model_zoo/openai-clip-vit-large-patch14",
+  "model_type": "mistral",
+  "num_attention_heads": 32,
+  "num_hidden_layers": 32,
+  "num_key_value_heads": 8,
+  "num_query": 32,
+  "rms_norm_eps": 1e-05,
+  "rope_theta": 10000.0,
+  "sliding_window": 4096,
+  "tie_word_embeddings": false,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.38.2",
+  "tune_mm_mlp_adapter": false,
+  "use_cache": false,
+  "use_mm_proj": true,
+  "vocab_size": 48384
+}

checkpoint-300/global_step300/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f9ba647e6820f714b07c06bf931419392ebe5ed2759842c9fd2429ded903d501
+size 6931872192

checkpoint-300/global_step300/mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:91249b58f5d288388c50bd94eedae7ce7b5e7db4a9a6ac18c01c9415b7afd83f
+size 1155894712

checkpoint-300/latest

@@ -0,0 +1 @@
+global_step300

checkpoint-300/non_lora_trainables.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f0e64ac33c32c2e6383611d3255e401592cb253e4eefe21a1ceaea1719f38730
+size 987789762

checkpoint-300/rng_state.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:62014b118bf0a56fb8931c6dcf66e897e597ebae4ca3836ba21f1c863467944e
+size 14244

checkpoint-300/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:54d8aec74ed1547670dc36c993b163ed9fe64c06c324dc0dbd57d2ec2238dceb
+size 1064

checkpoint-300/special_tokens_map.json

@@ -0,0 +1,24 @@
+{
+  "bos_token": {
+    "content": "<s>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "</s>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": "<unk>",
+  "unk_token": {
+    "content": "<unk>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

checkpoint-300/tokenizer.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4d88bdadaa2a065aa7c6e18a4b5999ce4c76cec14d9fea882102e7b4931d7ef0
+size 779539

checkpoint-300/tokenizer_config.json

@@ -0,0 +1,45 @@
+{
+  "add_bos_token": true,
+  "add_eos_token": false,
+  "add_prefix_space": true,
+  "added_tokens_decoder": {
+    "0": {
+      "content": "<unk>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "1": {
+      "content": "<s>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "2": {
+      "content": "</s>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "additional_special_tokens": [],
+  "bos_token": "<s>",
+  "chat_template": "{% if not add_generation_prompt is defined %}{% set add_generation_prompt = false %}{% endif %}{{ bos_token }}{% for message in messages %}{{'<|im_start|>' + message['role'] + '\n' + message['content'] + '</s>'}}{% endfor %}{% if add_generation_prompt %}{{ '<|im_start|>assistant\n' }}{% endif %}",
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "</s>",
+  "legacy": true,
+  "model_max_length": 4096,
+  "pad_token": "<unk>",
+  "padding_side": "right",
+  "sp_model_kwargs": {},
+  "spaces_between_special_tokens": false,
+  "tokenizer_class": "LlamaTokenizer",
+  "unk_token": "<unk>",
+  "use_default_system_prompt": false
+}

checkpoint-300/trainer_state.json

@@ -0,0 +1,2121 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 0.9236097748701174,
+  "eval_steps": 500,
+  "global_step": 300,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.0,
+      "grad_norm": 3.3905354112155957,
+      "learning_rate": 1e-05,
+      "loss": 1.5601,
+      "step": 1
+    },
+    {
+      "epoch": 0.01,
+      "grad_norm": 3.2069311238141744,
+      "learning_rate": 2e-05,
+      "loss": 1.5425,
+      "step": 2
+    },
+    {
+      "epoch": 0.01,
+      "grad_norm": 2.329876685543184,
+      "learning_rate": 3e-05,
+      "loss": 1.5518,
+      "step": 3
+    },
+    {
+      "epoch": 0.01,
+      "grad_norm": 1.3279617694098214,
+      "learning_rate": 4e-05,
+      "loss": 1.4658,
+      "step": 4
+    },
+    {
+      "epoch": 0.02,
+      "grad_norm": 1.3546782512787179,
+      "learning_rate": 5e-05,
+      "loss": 1.4736,
+      "step": 5
+    },
+    {
+      "epoch": 0.02,
+      "grad_norm": 1.6365619709572907,
+      "learning_rate": 6e-05,
+      "loss": 1.4756,
+      "step": 6
+    },
+    {
+      "epoch": 0.02,
+      "grad_norm": 1.2081088328094727,
+      "learning_rate": 7e-05,
+      "loss": 1.3752,
+      "step": 7
+    },
+    {
+      "epoch": 0.02,
+      "grad_norm": 1.0574546767027917,
+      "learning_rate": 8e-05,
+      "loss": 1.439,
+      "step": 8
+    },
+    {
+      "epoch": 0.03,
+      "grad_norm": 1.0012858470947548,
+      "learning_rate": 9e-05,
+      "loss": 1.4502,
+      "step": 9
+    },
+    {
+      "epoch": 0.03,
+      "grad_norm": 1.0136463608511321,
+      "learning_rate": 0.0001,
+      "loss": 1.396,
+      "step": 10
+    },
+    {
+      "epoch": 0.03,
+      "grad_norm": 0.9854986957079499,
+      "learning_rate": 9.999749748415981e-05,
+      "loss": 1.4634,
+      "step": 11
+    },
+    {
+      "epoch": 0.04,
+      "grad_norm": 0.9239702875144448,
+      "learning_rate": 9.998999018714263e-05,
+      "loss": 1.4185,
+      "step": 12
+    },
+    {
+      "epoch": 0.04,
+      "grad_norm": 0.9233186885946617,
+      "learning_rate": 9.997747886043367e-05,
+      "loss": 1.3691,
+      "step": 13
+    },
+    {
+      "epoch": 0.04,
+      "grad_norm": 0.8728080646228755,
+      "learning_rate": 9.995996475642466e-05,
+      "loss": 1.3569,
+      "step": 14
+    },
+    {
+      "epoch": 0.05,
+      "grad_norm": 0.8307436189120798,
+      "learning_rate": 9.99374496282885e-05,
+      "loss": 1.3667,
+      "step": 15
+    },
+    {
+      "epoch": 0.05,
+      "grad_norm": 0.7939681765489442,
+      "learning_rate": 9.990993572980378e-05,
+      "loss": 1.3589,
+      "step": 16
+    },
+    {
+      "epoch": 0.05,
+      "grad_norm": 0.7980526587202514,
+      "learning_rate": 9.987742581512918e-05,
+      "loss": 1.395,
+      "step": 17
+    },
+    {
+      "epoch": 0.06,
+      "grad_norm": 0.7207026961586376,
+      "learning_rate": 9.983992313852774e-05,
+      "loss": 1.3887,
+      "step": 18
+    },
+    {
+      "epoch": 0.06,
+      "grad_norm": 0.7871310923443939,
+      "learning_rate": 9.979743145404119e-05,
+      "loss": 1.3062,
+      "step": 19
+    },
+    {
+      "epoch": 0.06,
+      "grad_norm": 0.7246807835643384,
+      "learning_rate": 9.974995501511404e-05,
+      "loss": 1.4028,
+      "step": 20
+    },
+    {
+      "epoch": 0.06,
+      "grad_norm": 0.7206361737972257,
+      "learning_rate": 9.969749857416789e-05,
+      "loss": 1.3398,
+      "step": 21
+    },
+    {
+      "epoch": 0.07,
+      "grad_norm": 0.6960105391155793,
+      "learning_rate": 9.964006738212575e-05,
+      "loss": 1.3469,
+      "step": 22
+    },
+    {
+      "epoch": 0.07,
+      "grad_norm": 0.7204455050781057,
+      "learning_rate": 9.957766718788633e-05,
+      "loss": 1.3765,
+      "step": 23
+    },
+    {
+      "epoch": 0.07,
+      "grad_norm": 0.5812596677574942,
+      "learning_rate": 9.951030423774859e-05,
+      "loss": 1.3379,
+      "step": 24
+    },
+    {
+      "epoch": 0.08,
+      "grad_norm": 0.675904894861765,
+      "learning_rate": 9.943798527478651e-05,
+      "loss": 1.3789,
+      "step": 25
+    },
+    {
+      "epoch": 0.08,
+      "grad_norm": 0.6664199931252878,
+      "learning_rate": 9.936071753817415e-05,
+      "loss": 1.3535,
+      "step": 26
+    },
+    {
+      "epoch": 0.08,
+      "grad_norm": 0.7030363325377104,
+      "learning_rate": 9.927850876246088e-05,
+      "loss": 1.218,
+      "step": 27
+    },
+    {
+      "epoch": 0.09,
+      "grad_norm": 0.6185910243439269,
+      "learning_rate": 9.919136717679722e-05,
+      "loss": 1.3501,
+      "step": 28
+    },
+    {
+      "epoch": 0.09,
+      "grad_norm": 0.6270496500828271,
+      "learning_rate": 9.909930150411113e-05,
+      "loss": 1.3628,
+      "step": 29
+    },
+    {
+      "epoch": 0.09,
+      "grad_norm": 0.6021577621428597,
+      "learning_rate": 9.900232096023477e-05,
+      "loss": 1.3428,
+      "step": 30
+    },
+    {
+      "epoch": 0.1,
+      "grad_norm": 0.6760900210464318,
+      "learning_rate": 9.890043525298203e-05,
+      "loss": 1.3079,
+      "step": 31
+    },
+    {
+      "epoch": 0.1,
+      "grad_norm": 0.5940959503401227,
+      "learning_rate": 9.879365458117678e-05,
+      "loss": 1.291,
+      "step": 32
+    },
+    {
+      "epoch": 0.1,
+      "grad_norm": 0.5427094594745479,
+      "learning_rate": 9.868198963363189e-05,
+      "loss": 1.2893,
+      "step": 33
+    },
+    {
+      "epoch": 0.1,
+      "grad_norm": 0.5958301777472343,
+      "learning_rate": 9.856545158807938e-05,
+      "loss": 1.2363,
+      "step": 34
+    },
+    {
+      "epoch": 0.11,
+      "grad_norm": 0.5850214722817281,
+      "learning_rate": 9.844405211005146e-05,
+      "loss": 1.3154,
+      "step": 35
+    },
+    {
+      "epoch": 0.11,
+      "grad_norm": 0.6172288827652985,
+      "learning_rate": 9.831780335171279e-05,
+      "loss": 1.3101,
+      "step": 36
+    },
+    {
+      "epoch": 0.11,
+      "grad_norm": 0.5728085013088785,
+      "learning_rate": 9.818671795064404e-05,
+      "loss": 1.2817,
+      "step": 37
+    },
+    {
+      "epoch": 0.12,
+      "grad_norm": 0.5979780148525169,
+      "learning_rate": 9.805080902857699e-05,
+      "loss": 1.2493,
+      "step": 38
+    },
+    {
+      "epoch": 0.12,
+      "grad_norm": 0.537951216091522,
+      "learning_rate": 9.791009019008078e-05,
+      "loss": 1.2454,
+      "step": 39
+    },
+    {
+      "epoch": 0.12,
+      "grad_norm": 0.5879388223555754,
+      "learning_rate": 9.776457552120033e-05,
+      "loss": 1.2744,
+      "step": 40
+    },
+    {
+      "epoch": 0.13,
+      "grad_norm": 0.5828144715679795,
+      "learning_rate": 9.761427958804621e-05,
+      "loss": 1.2354,
+      "step": 41
+    },
+    {
+      "epoch": 0.13,
+      "grad_norm": 0.5640545987289305,
+      "learning_rate": 9.745921743533653e-05,
+      "loss": 1.3105,
+      "step": 42
+    },
+    {
+      "epoch": 0.13,
+      "grad_norm": 0.6077402891626962,
+      "learning_rate": 9.729940458489104e-05,
+      "loss": 1.3188,
+      "step": 43
+    },
+    {
+      "epoch": 0.14,
+      "grad_norm": 0.5649165442230879,
+      "learning_rate": 9.713485703407731e-05,
+      "loss": 1.2756,
+      "step": 44
+    },
+    {
+      "epoch": 0.14,
+      "grad_norm": 0.5751090024359722,
+      "learning_rate": 9.696559125420948e-05,
+      "loss": 1.2336,
+      "step": 45
+    },
+    {
+      "epoch": 0.14,
+      "grad_norm": 0.5415492855376868,
+      "learning_rate": 9.679162418889931e-05,
+      "loss": 1.2424,
+      "step": 46
+    },
+    {
+      "epoch": 0.14,
+      "grad_norm": 0.5742691889803935,
+      "learning_rate": 9.66129732523603e-05,
+      "loss": 1.2363,
+      "step": 47
+    },
+    {
+      "epoch": 0.15,
+      "grad_norm": 0.5763001671560306,
+      "learning_rate": 9.642965632766436e-05,
+      "loss": 1.2661,
+      "step": 48
+    },
+    {
+      "epoch": 0.15,
+      "grad_norm": 0.5940077836695673,
+      "learning_rate": 9.624169176495184e-05,
+      "loss": 1.2734,
+      "step": 49
+    },
+    {
+      "epoch": 0.15,
+      "grad_norm": 0.6005566973905004,
+      "learning_rate": 9.604909837959455e-05,
+      "loss": 1.2078,
+      "step": 50
+    },
+    {
+      "epoch": 0.16,
+      "grad_norm": 0.5429670344508561,
+      "learning_rate": 9.585189545031238e-05,
+      "loss": 1.2246,
+      "step": 51
+    },
+    {
+      "epoch": 0.16,
+      "grad_norm": 0.5871060174846529,
+      "learning_rate": 9.565010271724352e-05,
+      "loss": 1.3501,
+      "step": 52
+    },
+    {
+      "epoch": 0.16,
+      "grad_norm": 0.5907025473048154,
+      "learning_rate": 9.54437403799684e-05,
+      "loss": 1.2729,
+      "step": 53
+    },
+    {
+      "epoch": 0.17,
+      "grad_norm": 0.6516780485358238,
+      "learning_rate": 9.523282909548773e-05,
+      "loss": 1.2881,
+      "step": 54
+    },
+    {
+      "epoch": 0.17,
+      "grad_norm": 0.5670548168352986,
+      "learning_rate": 9.50173899761547e-05,
+      "loss": 1.2056,
+      "step": 55
+    },
+    {
+      "epoch": 0.17,
+      "grad_norm": 0.5390303783246289,
+      "learning_rate": 9.47974445875617e-05,
+      "loss": 1.2144,
+      "step": 56
+    },
+    {
+      "epoch": 0.18,
+      "grad_norm": 0.5646076772019145,
+      "learning_rate": 9.457301494638147e-05,
+      "loss": 1.2212,
+      "step": 57
+    },
+    {
+      "epoch": 0.18,
+      "grad_norm": 0.5612167041328082,
+      "learning_rate": 9.434412351816328e-05,
+      "loss": 1.2717,
+      "step": 58
+    },
+    {
+      "epoch": 0.18,
+      "grad_norm": 0.5263144027788276,
+      "learning_rate": 9.411079321508414e-05,
+      "loss": 1.2112,
+      "step": 59
+    },
+    {
+      "epoch": 0.18,
+      "grad_norm": 0.5356079649939604,
+      "learning_rate": 9.387304739365523e-05,
+      "loss": 1.2234,
+      "step": 60
+    },
+    {
+      "epoch": 0.19,
+      "grad_norm": 0.5615087563747105,
+      "learning_rate": 9.36309098523839e-05,
+      "loss": 1.2996,
+      "step": 61
+    },
+    {
+      "epoch": 0.19,
+      "grad_norm": 0.5314023973665841,
+      "learning_rate": 9.338440482939146e-05,
+      "loss": 1.1797,
+      "step": 62
+    },
+    {
+      "epoch": 0.19,
+      "grad_norm": 0.5539257915730327,
+      "learning_rate": 9.31335569999869e-05,
+      "loss": 1.2395,
+      "step": 63
+    },
+    {
+      "epoch": 0.2,
+      "grad_norm": 0.5648098416860655,
+      "learning_rate": 9.287839147419686e-05,
+      "loss": 1.2156,
+      "step": 64
+    },
+    {
+      "epoch": 0.2,
+      "grad_norm": 0.5865707594533118,
+      "learning_rate": 9.261893379425218e-05,
+      "loss": 1.2683,
+      "step": 65
+    },
+    {
+      "epoch": 0.2,
+      "grad_norm": 0.5440049015077518,
+      "learning_rate": 9.2355209932031e-05,
+      "loss": 1.1853,
+      "step": 66
+    },
+    {
+      "epoch": 0.21,
+      "grad_norm": 0.5440912007358205,
+      "learning_rate": 9.208724628645902e-05,
+      "loss": 1.207,
+      "step": 67
+    },
+    {
+      "epoch": 0.21,
+      "grad_norm": 0.5587040510710671,
+      "learning_rate": 9.181506968086697e-05,
+      "loss": 1.1768,
+      "step": 68
+    },
+    {
+      "epoch": 0.21,
+      "grad_norm": 0.5192024047252568,
+      "learning_rate": 9.153870736030548e-05,
+      "loss": 1.2427,
+      "step": 69
+    },
+    {
+      "epoch": 0.22,
+      "grad_norm": 0.528230045301502,
+      "learning_rate": 9.125818698881798e-05,
+      "loss": 1.2168,
+      "step": 70
+    },
+    {
+      "epoch": 0.22,
+      "grad_norm": 0.5562754356167662,
+      "learning_rate": 9.097353664667138e-05,
+      "loss": 1.2793,
+      "step": 71
+    },
+    {
+      "epoch": 0.22,
+      "grad_norm": 0.5415713371356312,
+      "learning_rate": 9.068478482754532e-05,
+      "loss": 1.2334,
+      "step": 72
+    },
+    {
+      "epoch": 0.22,
+      "grad_norm": 0.5683526535348254,
+      "learning_rate": 9.03919604356798e-05,
+      "loss": 1.2534,
+      "step": 73
+    },
+    {
+      "epoch": 0.23,
+      "grad_norm": 0.5396091696393609,
+      "learning_rate": 9.0095092782982e-05,
+      "loss": 1.2722,
+      "step": 74
+    },
+    {
+      "epoch": 0.23,
+      "grad_norm": 0.513589724983156,
+      "learning_rate": 8.979421158609206e-05,
+      "loss": 1.2083,
+      "step": 75
+    },
+    {
+      "epoch": 0.23,
+      "grad_norm": 0.5376666654796496,
+      "learning_rate": 8.948934696340843e-05,
+      "loss": 1.1931,
+      "step": 76
+    },
+    {
+      "epoch": 0.24,
+      "grad_norm": 0.5368664737517322,
+      "learning_rate": 8.918052943207298e-05,
+      "loss": 1.2202,
+      "step": 77
+    },
+    {
+      "epoch": 0.24,
+      "grad_norm": 0.5263413266320822,
+      "learning_rate": 8.886778990491631e-05,
+      "loss": 1.1553,
+      "step": 78
+    },
+    {
+      "epoch": 0.24,
+      "grad_norm": 0.5604541523029579,
+      "learning_rate": 8.85511596873632e-05,
+      "loss": 1.2964,
+      "step": 79
+    },
+    {
+      "epoch": 0.25,
+      "grad_norm": 0.5427504321680312,
+      "learning_rate": 8.823067047429907e-05,
+      "loss": 1.2896,
+      "step": 80
+    },
+    {
+      "epoch": 0.25,
+      "grad_norm": 0.5474275345159585,
+      "learning_rate": 8.790635434689721e-05,
+      "loss": 1.2612,
+      "step": 81
+    },
+    {
+      "epoch": 0.25,
+      "grad_norm": 0.5269654905955425,
+      "learning_rate": 8.757824376940746e-05,
+      "loss": 1.1821,
+      "step": 82
+    },
+    {
+      "epoch": 0.26,
+      "grad_norm": 0.518380839571753,
+      "learning_rate": 8.724637158590652e-05,
+      "loss": 1.2393,
+      "step": 83
+    },
+    {
+      "epoch": 0.26,
+      "grad_norm": 0.5158889921158187,
+      "learning_rate": 8.691077101701024e-05,
+      "loss": 1.2754,
+      "step": 84
+    },
+    {
+      "epoch": 0.26,
+      "grad_norm": 0.4935757204910804,
+      "learning_rate": 8.65714756565482e-05,
+      "loss": 1.2173,
+      "step": 85
+    },
+    {
+      "epoch": 0.26,
+      "grad_norm": 0.5653934280740491,
+      "learning_rate": 8.622851946820095e-05,
+      "loss": 1.2253,
+      "step": 86
+    },
+    {
+      "epoch": 0.27,
+      "grad_norm": 0.5250362872573285,
+      "learning_rate": 8.588193678210026e-05,
+      "loss": 1.1707,
+      "step": 87
+    },
+    {
+      "epoch": 0.27,
+      "grad_norm": 0.5446362423647451,
+      "learning_rate": 8.553176229139261e-05,
+      "loss": 1.1589,
+      "step": 88
+    },
+    {
+      "epoch": 0.27,
+      "grad_norm": 0.5142920460404157,
+      "learning_rate": 8.517803104876639e-05,
+      "loss": 1.1948,
+      "step": 89
+    },
+    {
+      "epoch": 0.28,
+      "grad_norm": 0.5317755291367837,
+      "learning_rate": 8.482077846294308e-05,
+      "loss": 1.1812,
+      "step": 90
+    },
+    {
+      "epoch": 0.28,
+      "grad_norm": 0.5147982642702206,
+      "learning_rate": 8.446004029513294e-05,
+      "loss": 1.2266,
+      "step": 91
+    },
+    {
+      "epoch": 0.28,
+      "grad_norm": 0.5096116915413985,
+      "learning_rate": 8.409585265545509e-05,
+      "loss": 1.179,
+      "step": 92
+    },
+    {
+      "epoch": 0.29,
+      "grad_norm": 0.49910238964612424,
+      "learning_rate": 8.372825199932304e-05,
+      "loss": 1.2485,
+      "step": 93
+    },
+    {
+      "epoch": 0.29,
+      "grad_norm": 0.5507929652198581,
+      "learning_rate": 8.335727512379534e-05,
+      "loss": 1.2388,
+      "step": 94
+    },
+    {
+      "epoch": 0.29,
+      "grad_norm": 0.5063227228339493,
+      "learning_rate": 8.298295916389234e-05,
+      "loss": 1.2415,
+      "step": 95
+    },
+    {
+      "epoch": 0.3,
+      "grad_norm": 0.48774944558422,
+      "learning_rate": 8.260534158887876e-05,
+      "loss": 1.1301,
+      "step": 96
+    },
+    {
+      "epoch": 0.3,
+      "grad_norm": 0.5131731261481731,
+      "learning_rate": 8.222446019851314e-05,
+      "loss": 1.2158,
+      "step": 97
+    },
+    {
+      "epoch": 0.3,
+      "grad_norm": 0.49068550228239954,
+      "learning_rate": 8.184035311926396e-05,
+      "loss": 1.1968,
+      "step": 98
+    },
+    {
+      "epoch": 0.3,
+      "grad_norm": 0.5288695654598018,
+      "learning_rate": 8.145305880049328e-05,
+      "loss": 1.2637,
+      "step": 99
+    },
+    {
+      "epoch": 0.31,
+      "grad_norm": 0.5479388159051014,
+      "learning_rate": 8.106261601060772e-05,
+      "loss": 1.3218,
+      "step": 100
+    },
+    {
+      "epoch": 0.31,
+      "grad_norm": 0.5621161336370604,
+      "learning_rate": 8.066906383317801e-05,
+      "loss": 1.1729,
+      "step": 101
+    },
+    {
+      "epoch": 0.31,
+      "grad_norm": 0.4924846556483893,
+      "learning_rate": 8.027244166302642e-05,
+      "loss": 1.1875,
+      "step": 102
+    },
+    {
+      "epoch": 0.32,
+      "grad_norm": 0.5037791405351364,
+      "learning_rate": 7.987278920228349e-05,
+      "loss": 1.2539,
+      "step": 103
+    },
+    {
+      "epoch": 0.32,
+      "grad_norm": 0.522728960991421,
+      "learning_rate": 7.947014645641379e-05,
+      "loss": 1.2217,
+      "step": 104
+    },
+    {
+      "epoch": 0.32,
+      "grad_norm": 0.5182027559256257,
+      "learning_rate": 7.906455373021129e-05,
+      "loss": 1.2188,
+      "step": 105
+    },
+    {
+      "epoch": 0.33,
+      "grad_norm": 0.5539472403444413,
+      "learning_rate": 7.865605162376486e-05,
+      "loss": 1.1509,
+      "step": 106
+    },
+    {
+      "epoch": 0.33,
+      "grad_norm": 0.5439005321857965,
+      "learning_rate": 7.824468102839419e-05,
+      "loss": 1.251,
+      "step": 107
+    },
+    {
+      "epoch": 0.33,
+      "grad_norm": 0.5093593984665626,
+      "learning_rate": 7.783048312255653e-05,
+      "loss": 1.229,
+      "step": 108
+    },
+    {
+      "epoch": 0.34,
+      "grad_norm": 0.5145253165856254,
+      "learning_rate": 7.741349936772469e-05,
+      "loss": 1.1824,
+      "step": 109
+    },
+    {
+      "epoch": 0.34,
+      "grad_norm": 0.5702243863746732,
+      "learning_rate": 7.699377150423672e-05,
+      "loss": 1.1582,
+      "step": 110
+    },
+    {
+      "epoch": 0.34,
+      "grad_norm": 0.5451183103162889,
+      "learning_rate": 7.65713415471177e-05,
+      "loss": 1.2122,
+      "step": 111
+    },
+    {
+      "epoch": 0.34,
+      "grad_norm": 0.5252426501201418,
+      "learning_rate": 7.614625178187402e-05,
+      "loss": 1.1833,
+      "step": 112
+    },
+    {
+      "epoch": 0.35,
+      "grad_norm": 0.48749296422821287,
+      "learning_rate": 7.571854476026048e-05,
+      "loss": 1.2705,
+      "step": 113
+    },
+    {
+      "epoch": 0.35,
+      "grad_norm": 0.502002191255964,
+      "learning_rate": 7.528826329602099e-05,
+      "loss": 1.2188,
+      "step": 114
+    },
+    {
+      "epoch": 0.35,
+      "grad_norm": 0.48918503246260797,
+      "learning_rate": 7.485545046060271e-05,
+      "loss": 1.1997,
+      "step": 115
+    },
+    {
+      "epoch": 0.36,
+      "grad_norm": 0.5000600132443745,
+      "learning_rate": 7.442014957884472e-05,
+      "loss": 1.2051,
+      "step": 116
+    },
+    {
+      "epoch": 0.36,
+      "grad_norm": 0.5463372631998543,
+      "learning_rate": 7.398240422464109e-05,
+      "loss": 1.2214,
+      "step": 117
+    },
+    {
+      "epoch": 0.36,
+      "grad_norm": 0.4961344121724833,
+      "learning_rate": 7.354225821657914e-05,
+      "loss": 1.208,
+      "step": 118
+    },
+    {
+      "epoch": 0.37,
+      "grad_norm": 0.5899278757904003,
+      "learning_rate": 7.309975561355312e-05,
+      "loss": 1.146,
+      "step": 119
+    },
+    {
+      "epoch": 0.37,
+      "grad_norm": 0.5023010444635025,
+      "learning_rate": 7.265494071035401e-05,
+      "loss": 1.1509,
+      "step": 120
+    },
+    {
+      "epoch": 0.37,
+      "grad_norm": 0.5098342526132932,
+      "learning_rate": 7.220785803323544e-05,
+      "loss": 1.1743,
+      "step": 121
+    },
+    {
+      "epoch": 0.38,
+      "grad_norm": 0.5181510224338065,
+      "learning_rate": 7.175855233545668e-05,
+      "loss": 1.1807,
+      "step": 122
+    },
+    {
+      "epoch": 0.38,
+      "grad_norm": 0.5126009147492929,
+      "learning_rate": 7.130706859280274e-05,
+      "loss": 1.1875,
+      "step": 123
+    },
+    {
+      "epoch": 0.38,
+      "grad_norm": 0.5694155609516423,
+      "learning_rate": 7.085345199908235e-05,
+      "loss": 1.1428,
+      "step": 124
+    },
+    {
+      "epoch": 0.38,
+      "grad_norm": 0.5320176914945207,
+      "learning_rate": 7.03977479616039e-05,
+      "loss": 1.219,
+      "step": 125
+    },
+    {
+      "epoch": 0.39,
+      "grad_norm": 0.5263190337396886,
+      "learning_rate": 6.994000209663036e-05,
+      "loss": 1.1709,
+      "step": 126
+    },
+    {
+      "epoch": 0.39,
+      "grad_norm": 0.5043242059465939,
+      "learning_rate": 6.948026022481279e-05,
+      "loss": 1.1968,
+      "step": 127
+    },
+    {
+      "epoch": 0.39,
+      "grad_norm": 0.5580562389957314,
+      "learning_rate": 6.901856836660386e-05,
+      "loss": 1.1494,
+      "step": 128
+    },
+    {
+      "epoch": 0.4,
+      "grad_norm": 0.495782343949626,
+      "learning_rate": 6.855497273765112e-05,
+      "loss": 1.2119,
+      "step": 129
+    },
+    {
+      "epoch": 0.4,
+      "grad_norm": 0.49510379613102157,
+      "learning_rate": 6.808951974417078e-05,
+      "loss": 1.2063,
+      "step": 130
+    },
+    {
+      "epoch": 0.4,
+      "grad_norm": 0.5536322418977505,
+      "learning_rate": 6.762225597830237e-05,
+      "loss": 1.2617,
+      "step": 131
+    },
+    {
+      "epoch": 0.41,
+      "grad_norm": 0.5668665640341709,
+      "learning_rate": 6.715322821344494e-05,
+      "loss": 1.1443,
+      "step": 132
+    },
+    {
+      "epoch": 0.41,
+      "grad_norm": 0.5083973254687169,
+      "learning_rate": 6.668248339957491e-05,
+      "loss": 1.2588,
+      "step": 133
+    },
+    {
+      "epoch": 0.41,
+      "grad_norm": 0.5188609075674584,
+      "learning_rate": 6.621006865854644e-05,
+      "loss": 1.1726,
+      "step": 134
+    },
+    {
+      "epoch": 0.42,
+      "grad_norm": 0.5083154005353754,
+      "learning_rate": 6.573603127937442e-05,
+      "loss": 1.1953,
+      "step": 135
+    },
+    {
+      "epoch": 0.42,
+      "grad_norm": 0.544802137981313,
+      "learning_rate": 6.526041871350086e-05,
+      "loss": 1.1853,
+      "step": 136
+    },
+    {
+      "epoch": 0.42,
+      "grad_norm": 0.4692918589408204,
+      "learning_rate": 6.478327857004495e-05,
+      "loss": 1.1267,
+      "step": 137
+    },
+    {
+      "epoch": 0.42,
+      "grad_norm": 0.504019443454899,
+      "learning_rate": 6.43046586110374e-05,
+      "loss": 1.2085,
+      "step": 138
+    },
+    {
+      "epoch": 0.43,
+      "grad_norm": 0.5071221863228176,
+      "learning_rate": 6.382460674663932e-05,
+      "loss": 1.2026,
+      "step": 139
+    },
+    {
+      "epoch": 0.43,
+      "grad_norm": 0.5146798822767533,
+      "learning_rate": 6.334317103034652e-05,
+      "loss": 1.1177,
+      "step": 140
+    },
+    {
+      "epoch": 0.43,
+      "grad_norm": 0.4996775842843673,
+      "learning_rate": 6.286039965417925e-05,
+      "loss": 1.1711,
+      "step": 141
+    },
+    {
+      "epoch": 0.44,
+      "grad_norm": 0.5288823443570614,
+      "learning_rate": 6.237634094385813e-05,
+      "loss": 1.1199,
+      "step": 142
+    },
+    {
+      "epoch": 0.44,
+      "grad_norm": 0.5214584824734313,
+      "learning_rate": 6.18910433539668e-05,
+      "loss": 1.1199,
+      "step": 143
+    },
+    {
+      "epoch": 0.44,
+      "grad_norm": 0.5287174552858058,
+      "learning_rate": 6.140455546310148e-05,
+      "loss": 1.2134,
+      "step": 144
+    },
+    {
+      "epoch": 0.45,
+      "grad_norm": 0.5254574455344191,
+      "learning_rate": 6.0916925969008275e-05,
+      "loss": 1.2578,
+      "step": 145
+    },
+    {
+      "epoch": 0.45,
+      "grad_norm": 0.5189177716169284,
+      "learning_rate": 6.042820368370854e-05,
+      "loss": 1.2236,
+      "step": 146
+    },
+    {
+      "epoch": 0.45,
+      "grad_norm": 0.5680190447979175,
+      "learning_rate": 5.993843752861266e-05,
+      "loss": 1.1443,
+      "step": 147
+    },
+    {
+      "epoch": 0.46,
+      "grad_norm": 0.471841513545942,
+      "learning_rate": 5.944767652962309e-05,
+      "loss": 1.1733,
+      "step": 148
+    },
+    {
+      "epoch": 0.46,
+      "grad_norm": 0.5131643851734395,
+      "learning_rate": 5.895596981222678e-05,
+      "loss": 1.1592,
+      "step": 149
+    },
+    {
+      "epoch": 0.46,
+      "grad_norm": 0.5069414300839418,
+      "learning_rate": 5.8463366596577706e-05,
+      "loss": 1.2537,
+      "step": 150
+    },
+    {
+      "epoch": 0.46,
+      "grad_norm": 0.5239372199223642,
+      "learning_rate": 5.796991619256985e-05,
+      "loss": 1.1743,
+      "step": 151
+    },
+    {
+      "epoch": 0.47,
+      "grad_norm": 0.5387326952860074,
+      "learning_rate": 5.747566799490132e-05,
+      "loss": 1.2075,
+      "step": 152
+    },
+    {
+      "epoch": 0.47,
+      "grad_norm": 0.5224076525892895,
+      "learning_rate": 5.6980671478129853e-05,
+      "loss": 1.2515,
+      "step": 153
+    },
+    {
+      "epoch": 0.47,
+      "grad_norm": 0.5609147216869792,
+      "learning_rate": 5.648497619172042e-05,
+      "loss": 1.1836,
+      "step": 154
+    },
+    {
+      "epoch": 0.48,
+      "grad_norm": 0.5465488763290158,
+      "learning_rate": 5.5988631755085264e-05,
+      "loss": 1.1433,
+      "step": 155
+    },
+    {
+      "epoch": 0.48,
+      "grad_norm": 0.5079660248232113,
+      "learning_rate": 5.549168785261698e-05,
+      "loss": 1.1812,
+      "step": 156
+    },
+    {
+      "epoch": 0.48,
+      "grad_norm": 0.5717154092705009,
+      "learning_rate": 5.499419422871506e-05,
+      "loss": 1.2068,
+      "step": 157
+    },
+    {
+      "epoch": 0.49,
+      "grad_norm": 0.5301493442832448,
+      "learning_rate": 5.4496200682806495e-05,
+      "loss": 1.2273,
+      "step": 158
+    },
+    {
+      "epoch": 0.49,
+      "grad_norm": 0.5098890110687092,
+      "learning_rate": 5.399775706436076e-05,
+      "loss": 1.2134,
+      "step": 159
+    },
+    {
+      "epoch": 0.49,
+      "grad_norm": 0.5365690403371399,
+      "learning_rate": 5.3498913267899864e-05,
+      "loss": 1.2051,
+      "step": 160
+    },
+    {
+      "epoch": 0.5,
+      "grad_norm": 0.5002287628666306,
+      "learning_rate": 5.299971922800391e-05,
+      "loss": 1.1255,
+      "step": 161
+    },
+    {
+      "epoch": 0.5,
+      "grad_norm": 0.5276270904967929,
+      "learning_rate": 5.250022491431259e-05,
+      "loss": 1.2124,
+      "step": 162
+    },
+    {
+      "epoch": 0.5,
+      "grad_norm": 0.5224515486361312,
+      "learning_rate": 5.200048032652318e-05,
+      "loss": 1.2559,
+      "step": 163
+    },
+    {
+      "epoch": 0.5,
+      "grad_norm": 0.49316666866194575,
+      "learning_rate": 5.150053548938557e-05,
+      "loss": 1.1421,
+      "step": 164
+    },
+    {
+      "epoch": 0.51,
+      "grad_norm": 0.5388196993705506,
+      "learning_rate": 5.100044044769472e-05,
+      "loss": 1.2017,
+      "step": 165
+    },
+    {
+      "epoch": 0.51,
+      "grad_norm": 0.4944161877205336,
+      "learning_rate": 5.0500245261281175e-05,
+      "loss": 1.1838,
+      "step": 166
+    },
+    {
+      "epoch": 0.51,
+      "grad_norm": 0.5075949586376647,
+      "learning_rate": 5e-05,
+      "loss": 1.2698,
+      "step": 167
+    },
+    {
+      "epoch": 0.52,
+      "grad_norm": 0.5288349964384436,
+      "learning_rate": 4.949975473871884e-05,
+      "loss": 1.2051,
+      "step": 168
+    },
+    {
+      "epoch": 0.52,
+      "grad_norm": 0.5298564213314307,
+      "learning_rate": 4.899955955230529e-05,
+      "loss": 1.1921,
+      "step": 169
+    },
+    {
+      "epoch": 0.52,
+      "grad_norm": 0.5006832164442894,
+      "learning_rate": 4.849946451061443e-05,
+      "loss": 1.0999,
+      "step": 170
+    },
+    {
+      "epoch": 0.53,
+      "grad_norm": 0.5180365116427443,
+      "learning_rate": 4.799951967347683e-05,
+      "loss": 1.1248,
+      "step": 171
+    },
+    {
+      "epoch": 0.53,
+      "grad_norm": 0.5085456414296031,
+      "learning_rate": 4.749977508568742e-05,
+      "loss": 1.0989,
+      "step": 172
+    },
+    {
+      "epoch": 0.53,
+      "grad_norm": 0.5062569741436713,
+      "learning_rate": 4.7000280771996104e-05,
+      "loss": 1.1851,
+      "step": 173
+    },
+    {
+      "epoch": 0.54,
+      "grad_norm": 0.49915526983252295,
+      "learning_rate": 4.650108673210015e-05,
+      "loss": 1.1565,
+      "step": 174
+    },
+    {
+      "epoch": 0.54,
+      "grad_norm": 0.530737295373366,
+      "learning_rate": 4.6002242935639254e-05,
+      "loss": 1.1367,
+      "step": 175
+    },
+    {
+      "epoch": 0.54,
+      "grad_norm": 0.5354173274123866,
+      "learning_rate": 4.550379931719351e-05,
+      "loss": 1.2231,
+      "step": 176
+    },
+    {
+      "epoch": 0.54,
+      "grad_norm": 0.4910072525351625,
+      "learning_rate": 4.500580577128495e-05,
+      "loss": 1.2354,
+      "step": 177
+    },
+    {
+      "epoch": 0.55,
+      "grad_norm": 0.5227332856552718,
+      "learning_rate": 4.4508312147383036e-05,
+      "loss": 1.2178,
+      "step": 178
+    },
+    {
+      "epoch": 0.55,
+      "grad_norm": 0.5288368214749951,
+      "learning_rate": 4.4011368244914755e-05,
+      "loss": 1.1731,
+      "step": 179
+    },
+    {
+      "epoch": 0.55,
+      "grad_norm": 0.5044952842565356,
+      "learning_rate": 4.3515023808279586e-05,
+      "loss": 1.1287,
+      "step": 180
+    },
+    {
+      "epoch": 0.56,
+      "grad_norm": 0.5325342803705855,
+      "learning_rate": 4.301932852187016e-05,
+      "loss": 1.2129,
+      "step": 181
+    },
+    {
+      "epoch": 0.56,
+      "grad_norm": 0.5415040662560654,
+      "learning_rate": 4.252433200509869e-05,
+      "loss": 1.3086,
+      "step": 182
+    },
+    {
+      "epoch": 0.56,
+      "grad_norm": 0.51763854212227,
+      "learning_rate": 4.203008380743016e-05,
+      "loss": 1.1216,
+      "step": 183
+    },
+    {
+      "epoch": 0.57,
+      "grad_norm": 0.5422430582487087,
+      "learning_rate": 4.1536633403422306e-05,
+      "loss": 1.1111,
+      "step": 184
+    },
+    {
+      "epoch": 0.57,
+      "grad_norm": 0.5022269460427503,
+      "learning_rate": 4.104403018777323e-05,
+      "loss": 1.1963,
+      "step": 185
+    },
+    {
+      "epoch": 0.57,
+      "grad_norm": 0.5631689643987351,
+      "learning_rate": 4.0552323470376916e-05,
+      "loss": 1.261,
+      "step": 186
+    },
+    {
+      "epoch": 0.58,
+      "grad_norm": 0.5083595167881586,
+      "learning_rate": 4.006156247138736e-05,
+      "loss": 1.2529,
+      "step": 187
+    },
+    {
+      "epoch": 0.58,
+      "grad_norm": 0.4982995201328196,
+      "learning_rate": 3.9571796316291476e-05,
+      "loss": 1.1938,
+      "step": 188
+    },
+    {
+      "epoch": 0.58,
+      "grad_norm": 0.5142407248745989,
+      "learning_rate": 3.908307403099174e-05,
+      "loss": 1.231,
+      "step": 189
+    },
+    {
+      "epoch": 0.58,
+      "grad_norm": 0.49845909814685674,
+      "learning_rate": 3.859544453689853e-05,
+      "loss": 1.1174,
+      "step": 190
+    },
+    {
+      "epoch": 0.59,
+      "grad_norm": 0.5044393346608876,
+      "learning_rate": 3.810895664603321e-05,
+      "loss": 1.1626,
+      "step": 191
+    },
+    {
+      "epoch": 0.59,
+      "grad_norm": 0.5028906716759755,
+      "learning_rate": 3.762365905614187e-05,
+      "loss": 1.1248,
+      "step": 192
+    },
+    {
+      "epoch": 0.59,
+      "grad_norm": 0.5052094751619233,
+      "learning_rate": 3.713960034582077e-05,
+      "loss": 1.1782,
+      "step": 193
+    },
+    {
+      "epoch": 0.6,
+      "grad_norm": 0.4927933528160678,
+      "learning_rate": 3.665682896965349e-05,
+      "loss": 1.1411,
+      "step": 194
+    },
+    {
+      "epoch": 0.6,
+      "grad_norm": 0.5194837061389329,
+      "learning_rate": 3.61753932533607e-05,
+      "loss": 1.1194,
+      "step": 195
+    },
+    {
+      "epoch": 0.6,
+      "grad_norm": 0.5003507430931593,
+      "learning_rate": 3.5695341388962614e-05,
+      "loss": 1.1438,
+      "step": 196
+    },
+    {
+      "epoch": 0.61,
+      "grad_norm": 0.4952909851844929,
+      "learning_rate": 3.521672142995506e-05,
+      "loss": 1.1274,
+      "step": 197
+    },
+    {
+      "epoch": 0.61,
+      "grad_norm": 0.515058338420701,
+      "learning_rate": 3.473958128649915e-05,
+      "loss": 1.2173,
+      "step": 198
+    },
+    {
+      "epoch": 0.61,
+      "grad_norm": 0.4649441022469789,
+      "learning_rate": 3.4263968720625594e-05,
+      "loss": 1.0901,
+      "step": 199
+    },
+    {
+      "epoch": 0.62,
+      "grad_norm": 0.5257495392411118,
+      "learning_rate": 3.378993134145356e-05,
+      "loss": 1.2124,
+      "step": 200
+    },
+    {
+      "epoch": 0.62,
+      "grad_norm": 0.4932641665110047,
+      "learning_rate": 3.33175166004251e-05,
+      "loss": 1.2024,
+      "step": 201
+    },
+    {
+      "epoch": 0.62,
+      "grad_norm": 0.47986279797474973,
+      "learning_rate": 3.284677178655507e-05,
+      "loss": 1.199,
+      "step": 202
+    },
+    {
+      "epoch": 0.62,
+      "grad_norm": 0.5192526802216362,
+      "learning_rate": 3.2377744021697644e-05,
+      "loss": 1.1348,
+      "step": 203
+    },
+    {
+      "epoch": 0.63,
+      "grad_norm": 0.49075397683891514,
+      "learning_rate": 3.1910480255829237e-05,
+      "loss": 1.209,
+      "step": 204
+    },
+    {
+      "epoch": 0.63,
+      "grad_norm": 0.5308583566356849,
+      "learning_rate": 3.1445027262348894e-05,
+      "loss": 1.1843,
+      "step": 205
+    },
+    {
+      "epoch": 0.63,
+      "grad_norm": 0.4754800140640908,
+      "learning_rate": 3.098143163339615e-05,
+      "loss": 1.0713,
+      "step": 206
+    },
+    {
+      "epoch": 0.64,
+      "grad_norm": 0.48744339587851926,
+      "learning_rate": 3.051973977518723e-05,
+      "loss": 1.1589,
+      "step": 207
+    },
+    {
+      "epoch": 0.64,
+      "grad_norm": 0.538169296667757,
+      "learning_rate": 3.0059997903369656e-05,
+      "loss": 1.1636,
+      "step": 208
+    },
+    {
+      "epoch": 0.64,
+      "grad_norm": 0.5174008006678488,
+      "learning_rate": 2.9602252038396093e-05,
+      "loss": 1.0088,
+      "step": 209
+    },
+    {
+      "epoch": 0.65,
+      "grad_norm": 0.5423637270068683,
+      "learning_rate": 2.914654800091768e-05,
+      "loss": 1.2637,
+      "step": 210
+    },
+    {
+      "epoch": 0.65,
+      "grad_norm": 0.49532340168296196,
+      "learning_rate": 2.8692931407197275e-05,
+      "loss": 1.1211,
+      "step": 211
+    },
+    {
+      "epoch": 0.65,
+      "grad_norm": 0.5223772948936137,
+      "learning_rate": 2.824144766454333e-05,
+      "loss": 1.1392,
+      "step": 212
+    },
+    {
+      "epoch": 0.66,
+      "grad_norm": 0.5301272448221841,
+      "learning_rate": 2.7792141966764568e-05,
+      "loss": 1.1663,
+      "step": 213
+    },
+    {
+      "epoch": 0.66,
+      "grad_norm": 0.5241453393589025,
+      "learning_rate": 2.7345059289646008e-05,
+      "loss": 1.1375,
+      "step": 214
+    },
+    {
+      "epoch": 0.66,
+      "grad_norm": 0.4946190439071369,
+      "learning_rate": 2.6900244386446904e-05,
+      "loss": 1.1843,
+      "step": 215
+    },
+    {
+      "epoch": 0.66,
+      "grad_norm": 0.4894503682723616,
+      "learning_rate": 2.6457741783420886e-05,
+      "loss": 1.1492,
+      "step": 216
+    },
+    {
+      "epoch": 0.67,
+      "grad_norm": 0.4852212167430057,
+      "learning_rate": 2.6017595775358926e-05,
+      "loss": 1.1592,
+      "step": 217
+    },
+    {
+      "epoch": 0.67,
+      "grad_norm": 0.5560815990000127,
+      "learning_rate": 2.5579850421155293e-05,
+      "loss": 1.1484,
+      "step": 218
+    },
+    {
+      "epoch": 0.67,
+      "grad_norm": 0.5035735515547093,
+      "learning_rate": 2.514454953939731e-05,
+      "loss": 1.1381,
+      "step": 219
+    },
+    {
+      "epoch": 0.68,
+      "grad_norm": 0.5057189594248047,
+      "learning_rate": 2.4711736703979018e-05,
+      "loss": 1.2036,
+      "step": 220
+    },
+    {
+      "epoch": 0.68,
+      "grad_norm": 0.5116012961964514,
+      "learning_rate": 2.428145523973952e-05,
+      "loss": 1.2212,
+      "step": 221
+    },
+    {
+      "epoch": 0.68,
+      "grad_norm": 0.48593835956939924,
+      "learning_rate": 2.3853748218126e-05,
+      "loss": 1.1505,
+      "step": 222
+    },
+    {
+      "epoch": 0.69,
+      "grad_norm": 0.5014137656412536,
+      "learning_rate": 2.342865845288232e-05,
+      "loss": 1.2246,
+      "step": 223
+    },
+    {
+      "epoch": 0.69,
+      "grad_norm": 0.5223070185696177,
+      "learning_rate": 2.3006228495763295e-05,
+      "loss": 1.158,
+      "step": 224
+    },
+    {
+      "epoch": 0.69,
+      "grad_norm": 0.5624122056478932,
+      "learning_rate": 2.258650063227533e-05,
+      "loss": 1.0566,
+      "step": 225
+    },
+    {
+      "epoch": 0.7,
+      "grad_norm": 0.5471967527021497,
+      "learning_rate": 2.2169516877443485e-05,
+      "loss": 1.1277,
+      "step": 226
+    },
+    {
+      "epoch": 0.7,
+      "grad_norm": 0.48337168913421796,
+      "learning_rate": 2.1755318971605826e-05,
+      "loss": 1.1956,
+      "step": 227
+    },
+    {
+      "epoch": 0.7,
+      "grad_norm": 0.5284626857313452,
+      "learning_rate": 2.1343948376235144e-05,
+      "loss": 1.196,
+      "step": 228
+    },
+    {
+      "epoch": 0.71,
+      "grad_norm": 0.5837213719579207,
+      "learning_rate": 2.0935446269788717e-05,
+      "loss": 1.0967,
+      "step": 229
+    },
+    {
+      "epoch": 0.71,
+      "grad_norm": 0.5165785735037434,
+      "learning_rate": 2.052985354358622e-05,
+      "loss": 1.2036,
+      "step": 230
+    },
+    {
+      "epoch": 0.71,
+      "grad_norm": 0.5536196580249753,
+      "learning_rate": 2.0127210797716524e-05,
+      "loss": 1.1572,
+      "step": 231
+    },
+    {
+      "epoch": 0.71,
+      "grad_norm": 0.4991948055619459,
+      "learning_rate": 1.9727558336973595e-05,
+      "loss": 1.2112,
+      "step": 232
+    },
+    {
+      "epoch": 0.72,
+      "grad_norm": 0.5271303549814744,
+      "learning_rate": 1.933093616682201e-05,
+      "loss": 1.0928,
+      "step": 233
+    },
+    {
+      "epoch": 0.72,
+      "grad_norm": 0.4755907436318383,
+      "learning_rate": 1.8937383989392294e-05,
+      "loss": 1.198,
+      "step": 234
+    },
+    {
+      "epoch": 0.72,
+      "grad_norm": 0.5071149662683123,
+      "learning_rate": 1.854694119950675e-05,
+      "loss": 1.1602,
+      "step": 235
+    },
+    {
+      "epoch": 0.73,
+      "grad_norm": 0.541369836697391,
+      "learning_rate": 1.8159646880736036e-05,
+      "loss": 1.1282,
+      "step": 236
+    },
+    {
+      "epoch": 0.73,
+      "grad_norm": 0.5083789491431415,
+      "learning_rate": 1.7775539801486867e-05,
+      "loss": 1.0872,
+      "step": 237
+    },
+    {
+      "epoch": 0.73,
+      "grad_norm": 0.5324743232408313,
+      "learning_rate": 1.739465841112125e-05,
+      "loss": 1.2222,
+      "step": 238
+    },
+    {
+      "epoch": 0.74,
+      "grad_norm": 0.48387142611649014,
+      "learning_rate": 1.701704083610768e-05,
+      "loss": 1.158,
+      "step": 239
+    },
+    {
+      "epoch": 0.74,
+      "grad_norm": 0.5141063576503123,
+      "learning_rate": 1.664272487620466e-05,
+      "loss": 1.1162,
+      "step": 240
+    },
+    {
+      "epoch": 0.74,
+      "grad_norm": 0.5408323267729829,
+      "learning_rate": 1.6271748000676983e-05,
+      "loss": 1.02,
+      "step": 241
+    },
+    {
+      "epoch": 0.75,
+      "grad_norm": 0.49627165191053163,
+      "learning_rate": 1.590414734454493e-05,
+      "loss": 1.0848,
+      "step": 242
+    },
+    {
+      "epoch": 0.75,
+      "grad_norm": 0.5376377474797022,
+      "learning_rate": 1.5539959704867085e-05,
+      "loss": 1.2019,
+      "step": 243
+    },
+    {
+      "epoch": 0.75,
+      "grad_norm": 0.49371506434647033,
+      "learning_rate": 1.517922153705692e-05,
+      "loss": 1.2151,
+      "step": 244
+    },
+    {
+      "epoch": 0.75,
+      "grad_norm": 0.523592301439241,
+      "learning_rate": 1.4821968951233638e-05,
+      "loss": 1.231,
+      "step": 245
+    },
+    {
+      "epoch": 0.76,
+      "grad_norm": 0.5074105674850246,
+      "learning_rate": 1.4468237708607397e-05,
+      "loss": 1.1519,
+      "step": 246
+    },
+    {
+      "epoch": 0.76,
+      "grad_norm": 0.5066768606133761,
+      "learning_rate": 1.4118063217899746e-05,
+      "loss": 1.1492,
+      "step": 247
+    },
+    {
+      "epoch": 0.76,
+      "grad_norm": 0.5116682098711993,
+      "learning_rate": 1.3771480531799052e-05,
+      "loss": 1.1477,
+      "step": 248
+    },
+    {
+      "epoch": 0.77,
+      "grad_norm": 0.5352752339451914,
+      "learning_rate": 1.342852434345181e-05,
+      "loss": 1.0964,
+      "step": 249
+    },
+    {
+      "epoch": 0.77,
+      "grad_norm": 0.5418770700535751,
+      "learning_rate": 1.308922898298977e-05,
+      "loss": 1.0803,
+      "step": 250
+    },
+    {
+      "epoch": 0.77,
+      "grad_norm": 0.5157514248366463,
+      "learning_rate": 1.2753628414093489e-05,
+      "loss": 1.1614,
+      "step": 251
+    },
+    {
+      "epoch": 0.78,
+      "grad_norm": 0.5351720244073139,
+      "learning_rate": 1.2421756230592534e-05,
+      "loss": 1.2224,
+      "step": 252
+    },
+    {
+      "epoch": 0.78,
+      "grad_norm": 0.5214450914678166,
+      "learning_rate": 1.2093645653102786e-05,
+      "loss": 1.2302,
+      "step": 253
+    },
+    {
+      "epoch": 0.78,
+      "grad_norm": 0.5409585268429792,
+      "learning_rate": 1.1769329525700935e-05,
+      "loss": 1.2104,
+      "step": 254
+    },
+    {
+      "epoch": 0.79,
+      "grad_norm": 0.5220300809084372,
+      "learning_rate": 1.144884031263681e-05,
+      "loss": 1.1047,
+      "step": 255
+    },
+    {
+      "epoch": 0.79,
+      "grad_norm": 0.5376568328092315,
+      "learning_rate": 1.1132210095083694e-05,
+      "loss": 1.2002,
+      "step": 256
+    },
+    {
+      "epoch": 0.79,
+      "grad_norm": 0.5414025185087213,
+      "learning_rate": 1.081947056792702e-05,
+      "loss": 1.1833,
+      "step": 257
+    },
+    {
+      "epoch": 0.79,
+      "grad_norm": 0.49383703911170834,
+      "learning_rate": 1.0510653036591583e-05,
+      "loss": 1.199,
+      "step": 258
+    },
+    {
+      "epoch": 0.8,
+      "grad_norm": 0.4916712820340113,
+      "learning_rate": 1.0205788413907952e-05,
+      "loss": 1.1292,
+      "step": 259
+    },
+    {
+      "epoch": 0.8,
+      "grad_norm": 0.5168351602558782,
+      "learning_rate": 9.904907217018e-06,
+      "loss": 1.2712,
+      "step": 260
+    },
+    {
+      "epoch": 0.8,
+      "grad_norm": 0.4856958512300068,
+      "learning_rate": 9.608039564320209e-06,
+      "loss": 1.1277,
+      "step": 261
+    },
+    {
+      "epoch": 0.81,
+      "grad_norm": 0.5106017400451819,
+      "learning_rate": 9.31521517245469e-06,
+      "loss": 1.1108,
+      "step": 262
+    },
+    {
+      "epoch": 0.81,
+      "grad_norm": 0.5036900746833531,
+      "learning_rate": 9.026463353328613e-06,
+      "loss": 1.1748,
+      "step": 263
+    },
+    {
+      "epoch": 0.81,
+      "grad_norm": 0.5033379794982934,
+      "learning_rate": 8.741813011182014e-06,
+      "loss": 1.1521,
+      "step": 264
+    },
+    {
+      "epoch": 0.82,
+      "grad_norm": 0.5332335505700774,
+      "learning_rate": 8.461292639694518e-06,
+      "loss": 1.0955,
+      "step": 265
+    },
+    {
+      "epoch": 0.82,
+      "grad_norm": 0.5125213937729919,
+      "learning_rate": 8.18493031913305e-06,
+      "loss": 1.1089,
+      "step": 266
+    },
+    {
+      "epoch": 0.82,
+      "grad_norm": 0.5148264028992737,
+      "learning_rate": 7.912753713540988e-06,
+      "loss": 1.177,
+      "step": 267
+    },
+    {
+      "epoch": 0.83,
+      "grad_norm": 0.5229543225675894,
+      "learning_rate": 7.644790067969005e-06,
+      "loss": 1.1646,
+      "step": 268
+    },
+    {
+      "epoch": 0.83,
+      "grad_norm": 0.541964343800654,
+      "learning_rate": 7.381066205747822e-06,
+      "loss": 1.1187,
+      "step": 269
+    },
+    {
+      "epoch": 0.83,
+      "grad_norm": 0.5266917236869202,
+      "learning_rate": 7.1216085258031414e-06,
+      "loss": 1.0901,
+      "step": 270
+    },
+    {
+      "epoch": 0.83,
+      "grad_norm": 0.49725116144459375,
+      "learning_rate": 6.866443000013117e-06,
+      "loss": 1.1526,
+      "step": 271
+    },
+    {
+      "epoch": 0.84,
+      "grad_norm": 0.5012370969649222,
+      "learning_rate": 6.6155951706085405e-06,
+      "loss": 1.1689,
+      "step": 272
+    },
+    {
+      "epoch": 0.84,
+      "grad_norm": 0.5129519814796674,
+      "learning_rate": 6.369090147616103e-06,
+      "loss": 1.1348,
+      "step": 273
+    },
+    {
+      "epoch": 0.84,
+      "grad_norm": 0.4926628244202811,
+      "learning_rate": 6.1269526063447765e-06,
+      "loss": 1.1538,
+      "step": 274
+    },
+    {
+      "epoch": 0.85,
+      "grad_norm": 0.5420926972187566,
+      "learning_rate": 5.889206784915862e-06,
+      "loss": 1.1484,
+      "step": 275
+    },
+    {
+      "epoch": 0.85,
+      "grad_norm": 0.4950562048422779,
+      "learning_rate": 5.6558764818367195e-06,
+      "loss": 1.2126,
+      "step": 276
+    },
+    {
+      "epoch": 0.85,
+      "grad_norm": 0.536783110803186,
+      "learning_rate": 5.426985053618544e-06,
+      "loss": 1.1301,
+      "step": 277
+    },
+    {
+      "epoch": 0.86,
+      "grad_norm": 0.5016583572718184,
+      "learning_rate": 5.2025554124383095e-06,
+      "loss": 1.1582,
+      "step": 278
+    },
+    {
+      "epoch": 0.86,
+      "grad_norm": 0.5263293678305013,
+      "learning_rate": 4.9826100238453135e-06,
+      "loss": 1.1577,
+      "step": 279
+    },
+    {
+      "epoch": 0.86,
+      "grad_norm": 0.5272007649322749,
+      "learning_rate": 4.767170904512292e-06,
+      "loss": 1.167,
+      "step": 280
+    },
+    {
+      "epoch": 0.87,
+      "grad_norm": 0.5063050835599103,
+      "learning_rate": 4.556259620031617e-06,
+      "loss": 1.1846,
+      "step": 281
+    },
+    {
+      "epoch": 0.87,
+      "grad_norm": 0.4910298010180289,
+      "learning_rate": 4.349897282756487e-06,
+      "loss": 1.1357,
+      "step": 282
+    },
+    {
+      "epoch": 0.87,
+      "grad_norm": 0.5379014923521229,
+      "learning_rate": 4.148104549687626e-06,
+      "loss": 1.0798,
+      "step": 283
+    },
+    {
+      "epoch": 0.87,
+      "grad_norm": 0.4865878079832316,
+      "learning_rate": 3.95090162040545e-06,
+      "loss": 1.1123,
+      "step": 284
+    },
+    {
+      "epoch": 0.88,
+      "grad_norm": 0.5282174531059731,
+      "learning_rate": 3.758308235048158e-06,
+      "loss": 1.1538,
+      "step": 285
+    },
+    {
+      "epoch": 0.88,
+      "grad_norm": 0.5010967709298223,
+      "learning_rate": 3.570343672335641e-06,
+      "loss": 1.1392,
+      "step": 286
+    },
+    {
+      "epoch": 0.88,
+      "grad_norm": 0.5189711142846094,
+      "learning_rate": 3.38702674763971e-06,
+      "loss": 1.1851,
+      "step": 287
+    },
+    {
+      "epoch": 0.89,
+      "grad_norm": 0.49664252429083205,
+      "learning_rate": 3.2083758111006945e-06,
+      "loss": 1.11,
+      "step": 288
+    },
+    {
+      "epoch": 0.89,
+      "grad_norm": 0.4859267407564821,
+      "learning_rate": 3.0344087457905346e-06,
+      "loss": 1.1719,
+      "step": 289
+    },
+    {
+      "epoch": 0.89,
+      "grad_norm": 0.5151543902141822,
+      "learning_rate": 2.86514296592269e-06,
+      "loss": 1.2102,
+      "step": 290
+    },
+    {
+      "epoch": 0.9,
+      "grad_norm": 0.5256996348656279,
+      "learning_rate": 2.7005954151089695e-06,
+      "loss": 1.1709,
+      "step": 291
+    },
+    {
+      "epoch": 0.9,
+      "grad_norm": 0.5102631445476464,
+      "learning_rate": 2.54078256466348e-06,
+      "loss": 1.2034,
+      "step": 292
+    },
+    {
+      "epoch": 0.9,
+      "grad_norm": 0.5093343267813872,
+      "learning_rate": 2.3857204119538014e-06,
+      "loss": 1.0969,
+      "step": 293
+    },
+    {
+      "epoch": 0.91,
+      "grad_norm": 0.5024898891939928,
+      "learning_rate": 2.2354244787996748e-06,
+      "loss": 1.1289,
+      "step": 294
+    },
+    {
+      "epoch": 0.91,
+      "grad_norm": 0.5279824056783474,
+      "learning_rate": 2.0899098099192273e-06,
+      "loss": 1.1672,
+      "step": 295
+    },
+    {
+      "epoch": 0.91,
+      "grad_norm": 0.5379712149279697,
+      "learning_rate": 1.9491909714230204e-06,
+      "loss": 1.1858,
+      "step": 296
+    },
+    {
+      "epoch": 0.91,
+      "grad_norm": 0.5436100119596193,
+      "learning_rate": 1.8132820493559521e-06,
+      "loss": 1.1201,
+      "step": 297
+    },
+    {
+      "epoch": 0.92,
+      "grad_norm": 0.5015550141769547,
+      "learning_rate": 1.6821966482872264e-06,
+      "loss": 1.1797,
+      "step": 298
+    },
+    {
+      "epoch": 0.92,
+      "grad_norm": 0.5119620808666944,
+      "learning_rate": 1.5559478899485447e-06,
+      "loss": 1.1493,
+      "step": 299
+    },
+    {
+      "epoch": 0.92,
+      "grad_norm": 0.4977128949907143,
+      "learning_rate": 1.434548411920622e-06,
+      "loss": 1.1821,
+      "step": 300
+    }
+  ],
+  "logging_steps": 1.0,
+  "max_steps": 324,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 1,
+  "save_steps": 20,
+  "total_flos": 2.736962932582646e+18,
+  "train_batch_size": 16,
+  "trial_name": null,
+  "trial_params": null
+}

checkpoint-300/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f20677a2ff6badcbfb3707fb2d9d9e2ee25e726ed7144276fc52ec843930e05c
+size 6584

checkpoint-300/zero_to_fp32.py

@@ -0,0 +1,592 @@
+#!/usr/bin/env python
+
+# Copyright (c) Microsoft Corporation.
+# SPDX-License-Identifier: Apache-2.0
+
+# DeepSpeed Team
+
+# This script extracts fp32 consolidated weights from a zero 1, 2 and 3 DeepSpeed checkpoints. It gets
+# copied into the top level checkpoint dir, so the user can easily do the conversion at any point in
+# the future. Once extracted, the weights don't require DeepSpeed and can be used in any
+# application.
+#
+# example: python zero_to_fp32.py . pytorch_model.bin
+
+import argparse
+import torch
+import glob
+import math
+import os
+import re
+from collections import OrderedDict
+from dataclasses import dataclass
+
+# while this script doesn't use deepspeed to recover data, since the checkpoints are pickled with
+# DeepSpeed data structures it has to be available in the current python environment.
+from deepspeed.utils import logger
+from deepspeed.checkpoint.constants import (DS_VERSION, OPTIMIZER_STATE_DICT, SINGLE_PARTITION_OF_FP32_GROUPS,
+                                            FP32_FLAT_GROUPS, ZERO_STAGE, PARTITION_COUNT, PARAM_SHAPES, BUFFER_NAMES,
+                                            FROZEN_PARAM_SHAPES, FROZEN_PARAM_FRAGMENTS)
+
+
+@dataclass
+class zero_model_state:
+    buffers: dict()
+    param_shapes: dict()
+    shared_params: list
+    ds_version: int
+    frozen_param_shapes: dict()
+    frozen_param_fragments: dict()
+
+
+debug = 0
+
+# load to cpu
+device = torch.device('cpu')
+
+
+def atoi(text):
+    return int(text) if text.isdigit() else text
+
+
+def natural_keys(text):
+    '''
+    alist.sort(key=natural_keys) sorts in human order
+    http://nedbatchelder.com/blog/200712/human_sorting.html
+    (See Toothy's implementation in the comments)
+    '''
+    return [atoi(c) for c in re.split(r'(\d+)', text)]
+
+
+def get_model_state_file(checkpoint_dir, zero_stage):
+    if not os.path.isdir(checkpoint_dir):
+        raise FileNotFoundError(f"Directory '{checkpoint_dir}' doesn't exist")
+
+    # there should be only one file
+    if zero_stage <= 2:
+        file = os.path.join(checkpoint_dir, "mp_rank_00_model_states.pt")
+    elif zero_stage == 3:
+        file = os.path.join(checkpoint_dir, "zero_pp_rank_0_mp_rank_00_model_states.pt")
+
+    if not os.path.exists(file):
+        raise FileNotFoundError(f"can't find model states file at '{file}'")
+
+    return file
+
+
+def get_checkpoint_files(checkpoint_dir, glob_pattern):
+    # XXX: need to test that this simple glob rule works for multi-node setup too
+    ckpt_files = sorted(glob.glob(os.path.join(checkpoint_dir, glob_pattern)), key=natural_keys)
+
+    if len(ckpt_files) == 0:
+        raise FileNotFoundError(f"can't find {glob_pattern} files in directory '{checkpoint_dir}'")
+
+    return ckpt_files
+
+
+def get_optim_files(checkpoint_dir):
+    return get_checkpoint_files(checkpoint_dir, "*_optim_states.pt")
+
+
+def get_model_state_files(checkpoint_dir):
+    return get_checkpoint_files(checkpoint_dir, "*_model_states.pt")
+
+
+def parse_model_states(files):
+    zero_model_states = []
+    for file in files:
+        state_dict = torch.load(file, map_location=device)
+
+        if BUFFER_NAMES not in state_dict:
+            raise ValueError(f"{file} is not a model state checkpoint")
+        buffer_names = state_dict[BUFFER_NAMES]
+        if debug:
+            print("Found buffers:", buffer_names)
+
+        # recover just the buffers while restoring them to fp32 if they were saved in fp16
+        buffers = {k: v.float() for k, v in state_dict["module"].items() if k in buffer_names}
+        param_shapes = state_dict[PARAM_SHAPES]
+
+        # collect parameters that are included in param_shapes
+        param_names = []
+        for s in param_shapes:
+            for name in s.keys():
+                param_names.append(name)
+
+        # update with frozen parameters
+        frozen_param_shapes = state_dict.get(FROZEN_PARAM_SHAPES, None)
+        if frozen_param_shapes is not None:
+            if debug:
+                print(f"Found frozen_param_shapes: {frozen_param_shapes}")
+            param_names += list(frozen_param_shapes.keys())
+
+        # handle shared params
+        shared_params = [[k, v] for k, v in state_dict["shared_params"].items()]
+
+        ds_version = state_dict.get(DS_VERSION, None)
+
+        frozen_param_fragments = state_dict.get(FROZEN_PARAM_FRAGMENTS, None)
+
+        z_model_state = zero_model_state(buffers=buffers,
+                                         param_shapes=param_shapes,
+                                         shared_params=shared_params,
+                                         ds_version=ds_version,
+                                         frozen_param_shapes=frozen_param_shapes,
+                                         frozen_param_fragments=frozen_param_fragments)
+        zero_model_states.append(z_model_state)
+
+    return zero_model_states
+
+
+def parse_optim_states(files, ds_checkpoint_dir):
+
+    total_files = len(files)
+    state_dicts = []
+    for f in files:
+        state_dict = torch.load(f, map_location=device)
+        # immediately discard the potentially huge 2 optimizer states as we only care for fp32 master weights
+        # and also handle the case where it was already removed by another helper script
+        state_dict["optimizer_state_dict"].pop("optimizer_state_dict", None)
+        state_dicts.append(state_dict)
+
+    if not ZERO_STAGE in state_dicts[0][OPTIMIZER_STATE_DICT]:
+        raise ValueError(f"{files[0]} is not a zero checkpoint")
+    zero_stage = state_dicts[0][OPTIMIZER_STATE_DICT][ZERO_STAGE]
+    world_size = state_dicts[0][OPTIMIZER_STATE_DICT][PARTITION_COUNT]
+
+    # For ZeRO-2 each param group can have different partition_count as data parallelism for expert
+    # parameters can be different from data parallelism for non-expert parameters. So we can just
+    # use the max of the partition_count to get the dp world_size.
+
+    if type(world_size) is list:
+        world_size = max(world_size)
+
+    if world_size != total_files:
+        raise ValueError(
+            f"Expected {world_size} of '*_optim_states.pt' under '{ds_checkpoint_dir}' but found {total_files} files. "
+            "Possibly due to an overwrite of an old checkpoint, or a checkpoint didn't get saved by one or more processes."
+        )
+
+    # the groups are named differently in each stage
+    if zero_stage <= 2:
+        fp32_groups_key = SINGLE_PARTITION_OF_FP32_GROUPS
+    elif zero_stage == 3:
+        fp32_groups_key = FP32_FLAT_GROUPS
+    else:
+        raise ValueError(f"unknown zero stage {zero_stage}")
+
+    if zero_stage <= 2:
+        fp32_flat_groups = [state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key] for i in range(len(state_dicts))]
+    elif zero_stage == 3:
+        # if there is more than one param group, there will be multiple flattened tensors - one
+        # flattened tensor per group - for simplicity merge them into a single tensor
+        #
+        # XXX: could make the script more memory efficient for when there are multiple groups - it
+        # will require matching the sub-lists of param_shapes for each param group flattened tensor
+
+        fp32_flat_groups = [
+            torch.cat(state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key], 0) for i in range(len(state_dicts))
+        ]
+
+    return zero_stage, world_size, fp32_flat_groups
+
+
+def _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir):
+    """
+    Returns fp32 state_dict reconstructed from ds checkpoint
+
+    Args:
+        - ``ds_checkpoint_dir``: path to the deepspeed checkpoint folder (where the optimizer files are)
+
+    """
+    print(f"Processing zero checkpoint '{ds_checkpoint_dir}'")
+
+    optim_files = get_optim_files(ds_checkpoint_dir)
+    zero_stage, world_size, fp32_flat_groups = parse_optim_states(optim_files, ds_checkpoint_dir)
+    print(f"Detected checkpoint of type zero stage {zero_stage}, world_size: {world_size}")
+
+    model_files = get_model_state_files(ds_checkpoint_dir)
+
+    zero_model_states = parse_model_states(model_files)
+    print(f'Parsing checkpoint created by deepspeed=={zero_model_states[0].ds_version}')
+
+    if zero_stage <= 2:
+        return _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states)
+    elif zero_stage == 3:
+        return _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states)
+
+
+def _zero2_merge_frozen_params(state_dict, zero_model_states):
+    if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
+        return
+
+    frozen_param_shapes = zero_model_states[0].frozen_param_shapes
+    frozen_param_fragments = zero_model_states[0].frozen_param_fragments
+
+    if debug:
+        num_elem = sum(s.numel() for s in frozen_param_shapes.values())
+        print(f'rank 0: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
+
+        wanted_params = len(frozen_param_shapes)
+        wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
+        avail_numel = sum([p.numel() for p in frozen_param_fragments.values()])
+        print(f'Frozen params: Have {avail_numel} numels to process.')
+        print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
+
+    total_params = 0
+    total_numel = 0
+    for name, shape in frozen_param_shapes.items():
+        total_params += 1
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+
+        state_dict[name] = frozen_param_fragments[name]
+
+        if debug:
+            print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
+
+    print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _has_callable(obj, fn):
+    attr = getattr(obj, fn, None)
+    return callable(attr)
+
+
+def _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
+    param_shapes = zero_model_states[0].param_shapes
+
+    # Reconstruction protocol:
+    #
+    # XXX: document this
+
+    if debug:
+        for i in range(world_size):
+            for j in range(len(fp32_flat_groups[0])):
+                print(f"{FP32_FLAT_GROUPS}[{i}][{j}].shape={fp32_flat_groups[i][j].shape}")
+
+    # XXX: memory usage doubles here (zero2)
+    num_param_groups = len(fp32_flat_groups[0])
+    merged_single_partition_of_fp32_groups = []
+    for i in range(num_param_groups):
+        merged_partitions = [sd[i] for sd in fp32_flat_groups]
+        full_single_fp32_vector = torch.cat(merged_partitions, 0)
+        merged_single_partition_of_fp32_groups.append(full_single_fp32_vector)
+    avail_numel = sum(
+        [full_single_fp32_vector.numel() for full_single_fp32_vector in merged_single_partition_of_fp32_groups])
+
+    if debug:
+        wanted_params = sum([len(shapes) for shapes in param_shapes])
+        wanted_numel = sum([sum(shape.numel() for shape in shapes.values()) for shapes in param_shapes])
+        # not asserting if there is a mismatch due to possible padding
+        print(f"Have {avail_numel} numels to process.")
+        print(f"Need {wanted_numel} numels in {wanted_params} params.")
+
+    # params
+    # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
+    # out-of-core computing solution
+    total_numel = 0
+    total_params = 0
+    for shapes, full_single_fp32_vector in zip(param_shapes, merged_single_partition_of_fp32_groups):
+        offset = 0
+        avail_numel = full_single_fp32_vector.numel()
+        for name, shape in shapes.items():
+
+            unpartitioned_numel = shape.numel() if _has_callable(shape, 'numel') else math.prod(shape)
+            total_numel += unpartitioned_numel
+            total_params += 1
+
+            if debug:
+                print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
+            state_dict[name] = full_single_fp32_vector.narrow(0, offset, unpartitioned_numel).view(shape)
+            offset += unpartitioned_numel
+
+        # Z2 started to align to 2*world_size to improve nccl performance. Therefore both offset and
+        # avail_numel can differ by anywhere between 0..2*world_size. Due to two unrelated complex
+        # paddings performed in the code it's almost impossible to predict the exact numbers w/o the
+        # live optimizer object, so we are checking that the numbers are within the right range
+        align_to = 2 * world_size
+
+        def zero2_align(x):
+            return align_to * math.ceil(x / align_to)
+
+        if debug:
+            print(f"original offset={offset}, avail_numel={avail_numel}")
+
+        offset = zero2_align(offset)
+        avail_numel = zero2_align(avail_numel)
+
+        if debug:
+            print(f"aligned  offset={offset}, avail_numel={avail_numel}")
+
+        # Sanity check
+        if offset != avail_numel:
+            raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
+
+    print(f"Reconstructed fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states):
+    state_dict = OrderedDict()
+
+    # buffers
+    buffers = zero_model_states[0].buffers
+    state_dict.update(buffers)
+    if debug:
+        print(f"added {len(buffers)} buffers")
+
+    _zero2_merge_frozen_params(state_dict, zero_model_states)
+
+    _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
+
+    # recover shared parameters
+    for pair in zero_model_states[0].shared_params:
+        if pair[1] in state_dict:
+            state_dict[pair[0]] = state_dict[pair[1]]
+
+    return state_dict
+
+
+def zero3_partitioned_param_info(unpartitioned_numel, world_size):
+    remainder = unpartitioned_numel % world_size
+    padding_numel = (world_size - remainder) if remainder else 0
+    partitioned_numel = math.ceil(unpartitioned_numel / world_size)
+    return partitioned_numel, padding_numel
+
+
+def _zero3_merge_frozen_params(state_dict, world_size, zero_model_states):
+    if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
+        return
+
+    if debug:
+        for i in range(world_size):
+            num_elem = sum(s.numel() for s in zero_model_states[i].frozen_param_fragments.values())
+            print(f'rank {i}: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
+
+        frozen_param_shapes = zero_model_states[0].frozen_param_shapes
+        wanted_params = len(frozen_param_shapes)
+        wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
+        avail_numel = sum([p.numel() for p in zero_model_states[0].frozen_param_fragments.values()]) * world_size
+        print(f'Frozen params: Have {avail_numel} numels to process.')
+        print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
+
+    total_params = 0
+    total_numel = 0
+    for name, shape in zero_model_states[0].frozen_param_shapes.items():
+        total_params += 1
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+
+        param_frags = tuple(model_state.frozen_param_fragments[name] for model_state in zero_model_states)
+        state_dict[name] = torch.cat(param_frags, 0).narrow(0, 0, unpartitioned_numel).view(shape)
+
+        partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
+
+        if debug:
+            print(
+                f"Frozen params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
+            )
+
+    print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
+    param_shapes = zero_model_states[0].param_shapes
+    avail_numel = fp32_flat_groups[0].numel() * world_size
+    # Reconstruction protocol: For zero3 we need to zip the partitions together at boundary of each
+    # param, re-consolidating each param, while dealing with padding if any
+
+    # merge list of dicts, preserving order
+    param_shapes = {k: v for d in param_shapes for k, v in d.items()}
+
+    if debug:
+        for i in range(world_size):
+            print(f"{FP32_FLAT_GROUPS}[{i}].shape={fp32_flat_groups[i].shape}")
+
+        wanted_params = len(param_shapes)
+        wanted_numel = sum(shape.numel() for shape in param_shapes.values())
+        # not asserting if there is a mismatch due to possible padding
+        avail_numel = fp32_flat_groups[0].numel() * world_size
+        print(f"Trainable params: Have {avail_numel} numels to process.")
+        print(f"Trainable params: Need {wanted_numel} numels in {wanted_params} params.")
+
+    # params
+    # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
+    # out-of-core computing solution
+    offset = 0
+    total_numel = 0
+    total_params = 0
+    for name, shape in param_shapes.items():
+
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+        total_params += 1
+
+        partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
+
+        if debug:
+            print(
+                f"Trainable params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
+            )
+
+        # XXX: memory usage doubles here
+        state_dict[name] = torch.cat(
+            tuple(fp32_flat_groups[i].narrow(0, offset, partitioned_numel) for i in range(world_size)),
+            0).narrow(0, 0, unpartitioned_numel).view(shape)
+        offset += partitioned_numel
+
+    offset *= world_size
+
+    # Sanity check
+    if offset != avail_numel:
+        raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
+
+    print(f"Reconstructed Trainable fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states):
+    state_dict = OrderedDict()
+
+    # buffers
+    buffers = zero_model_states[0].buffers
+    state_dict.update(buffers)
+    if debug:
+        print(f"added {len(buffers)} buffers")
+
+    _zero3_merge_frozen_params(state_dict, world_size, zero_model_states)
+
+    _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
+
+    # recover shared parameters
+    for pair in zero_model_states[0].shared_params:
+        if pair[1] in state_dict:
+            state_dict[pair[0]] = state_dict[pair[1]]
+
+    return state_dict
+
+
+def get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag=None):
+    """
+    Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated state_dict that can be loaded with
+    ``load_state_dict()`` and used for training without DeepSpeed or shared with others, for example
+    via a model hub.
+
+    Args:
+        - ``checkpoint_dir``: path to the desired checkpoint folder
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in 'latest' file. e.g., ``global_step14``
+
+    Returns:
+        - pytorch ``state_dict``
+
+    Note: this approach may not work if your application doesn't have sufficient free CPU memory and
+    you may need to use the offline approach using the ``zero_to_fp32.py`` script that is saved with
+    the checkpoint.
+
+    A typical usage might be ::
+
+        from deepspeed.utils.zero_to_fp32 import get_fp32_state_dict_from_zero_checkpoint
+        # do the training and checkpoint saving
+        state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir) # already on cpu
+        model = model.cpu() # move to cpu
+        model.load_state_dict(state_dict)
+        # submit to model hub or save the model to share with others
+
+    In this example the ``model`` will no longer be usable in the deepspeed context of the same
+    application. i.e. you will need to re-initialize the deepspeed engine, since
+    ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
+
+    If you want it all done for you, use ``load_state_dict_from_zero_checkpoint`` instead.
+
+    """
+    if tag is None:
+        latest_path = os.path.join(checkpoint_dir, 'latest')
+        if os.path.isfile(latest_path):
+            with open(latest_path, 'r') as fd:
+                tag = fd.read().strip()
+        else:
+            raise ValueError(f"Unable to find 'latest' file at {latest_path}")
+
+    ds_checkpoint_dir = os.path.join(checkpoint_dir, tag)
+
+    if not os.path.isdir(ds_checkpoint_dir):
+        raise FileNotFoundError(f"Directory '{ds_checkpoint_dir}' doesn't exist")
+
+    return _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir)
+
+
+def convert_zero_checkpoint_to_fp32_state_dict(checkpoint_dir, output_file, tag=None):
+    """
+    Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` file that can be
+    loaded with ``torch.load(file)`` + ``load_state_dict()`` and used for training without DeepSpeed.
+
+    Args:
+        - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
+        - ``output_file``: path to the pytorch fp32 state_dict output file (e.g. path/pytorch_model.bin)
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
+    """
+
+    state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag)
+    print(f"Saving fp32 state dict to {output_file}")
+    torch.save(state_dict, output_file)
+
+
+def load_state_dict_from_zero_checkpoint(model, checkpoint_dir, tag=None):
+    """
+    1. Put the provided model to cpu
+    2. Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict``
+    3. Load it into the provided model
+
+    Args:
+        - ``model``: the model object to update
+        - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
+
+    Returns:
+        - ``model`: modified model
+
+    Make sure you have plenty of CPU memory available before you call this function. If you don't
+    have enough use the ``zero_to_fp32.py`` utility to do the conversion. You will find it
+    conveniently placed for you in the checkpoint folder.
+
+    A typical usage might be ::
+
+        from deepspeed.utils.zero_to_fp32 import load_state_dict_from_zero_checkpoint
+        model = load_state_dict_from_zero_checkpoint(trainer.model, checkpoint_dir)
+        # submit to model hub or save the model to share with others
+
+    Note, that once this was run, the ``model`` will no longer be usable in the deepspeed context
+    of the same application. i.e. you will need to re-initialize the deepspeed engine, since
+    ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
+
+    """
+    logger.info(f"Extracting fp32 weights")
+    state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag)
+
+    logger.info(f"Overwriting model with fp32 weights")
+    model = model.cpu()
+    model.load_state_dict(state_dict, strict=False)
+
+    return model
+
+
+if __name__ == "__main__":
+
+    parser = argparse.ArgumentParser()
+    parser.add_argument("checkpoint_dir",
+                        type=str,
+                        help="path to the desired checkpoint folder, e.g., path/checkpoint-12")
+    parser.add_argument(
+        "output_file",
+        type=str,
+        help="path to the pytorch fp32 state_dict output file (e.g. path/checkpoint-12/pytorch_model.bin)")
+    parser.add_argument("-t",
+                        "--tag",
+                        type=str,
+                        default=None,
+                        help="checkpoint tag used as a unique identifier for checkpoint. e.g., global_step1")
+    parser.add_argument("-d", "--debug", action='store_true', help="enable debug")
+    args = parser.parse_args()
+
+    debug = args.debug
+
+    convert_zero_checkpoint_to_fp32_state_dict(args.checkpoint_dir, args.output_file, tag=args.tag)

config.json

@@ -0,0 +1,43 @@
+{
+  "_name_or_path": "/model_zoo/Vivid-7B-base",
+  "architectures": [
+    "MistralForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "bert_type": "raw_bert_layer:12",
+  "bos_token_id": 1,
+  "compress_type": "mean",
+  "eos_token_id": 2,
+  "freeze_mm_mlp_adapter": false,
+  "hidden_act": "silu",
+  "hidden_size": 4096,
+  "image_aspect_ratio": "pad",
+  "image_grid_pinpoints": null,
+  "image_processor": "./llamavid/processor/clip-patch14-224",
+  "initializer_range": 0.02,
+  "intermediate_size": 14336,
+  "max_position_embeddings": 32768,
+  "max_token": 4096,
+  "mm_hidden_size": 1024,
+  "mm_projector_type": "mlp2x_gelu",
+  "mm_use_im_patch_token": false,
+  "mm_use_im_start_end": false,
+  "mm_vision_select_feature": "patch",
+  "mm_vision_select_layer": -2,
+  "mm_vision_tower": "/model_zoo/openai-clip-vit-large-patch14",
+  "model_type": "mistral",
+  "num_attention_heads": 32,
+  "num_hidden_layers": 32,
+  "num_key_value_heads": 8,
+  "num_query": 32,
+  "rms_norm_eps": 1e-05,
+  "rope_theta": 10000.0,
+  "sliding_window": 4096,
+  "tie_word_embeddings": false,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.38.2",
+  "tune_mm_mlp_adapter": false,
+  "use_cache": true,
+  "use_mm_proj": true,
+  "vocab_size": 48384
+}

non_lora_trainables.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d592844bd597e67cb6ac1906125c4b517602f9faf6be80ae7422d1088f57c54c
+size 987789762

trainer_state.json

@@ -0,0 +1,2298 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 0.9974985568597268,
+  "eval_steps": 500,
+  "global_step": 324,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.0,
+      "grad_norm": 3.3905354112155957,
+      "learning_rate": 1e-05,
+      "loss": 1.5601,
+      "step": 1
+    },
+    {
+      "epoch": 0.01,
+      "grad_norm": 3.2069311238141744,
+      "learning_rate": 2e-05,
+      "loss": 1.5425,
+      "step": 2
+    },
+    {
+      "epoch": 0.01,
+      "grad_norm": 2.329876685543184,
+      "learning_rate": 3e-05,
+      "loss": 1.5518,
+      "step": 3
+    },
+    {
+      "epoch": 0.01,
+      "grad_norm": 1.3279617694098214,
+      "learning_rate": 4e-05,
+      "loss": 1.4658,
+      "step": 4
+    },
+    {
+      "epoch": 0.02,
+      "grad_norm": 1.3546782512787179,
+      "learning_rate": 5e-05,
+      "loss": 1.4736,
+      "step": 5
+    },
+    {
+      "epoch": 0.02,
+      "grad_norm": 1.6365619709572907,
+      "learning_rate": 6e-05,
+      "loss": 1.4756,
+      "step": 6
+    },
+    {
+      "epoch": 0.02,
+      "grad_norm": 1.2081088328094727,
+      "learning_rate": 7e-05,
+      "loss": 1.3752,
+      "step": 7
+    },
+    {
+      "epoch": 0.02,
+      "grad_norm": 1.0574546767027917,
+      "learning_rate": 8e-05,
+      "loss": 1.439,
+      "step": 8
+    },
+    {
+      "epoch": 0.03,
+      "grad_norm": 1.0012858470947548,
+      "learning_rate": 9e-05,
+      "loss": 1.4502,
+      "step": 9
+    },
+    {
+      "epoch": 0.03,
+      "grad_norm": 1.0136463608511321,
+      "learning_rate": 0.0001,
+      "loss": 1.396,
+      "step": 10
+    },
+    {
+      "epoch": 0.03,
+      "grad_norm": 0.9854986957079499,
+      "learning_rate": 9.999749748415981e-05,
+      "loss": 1.4634,
+      "step": 11
+    },
+    {
+      "epoch": 0.04,
+      "grad_norm": 0.9239702875144448,
+      "learning_rate": 9.998999018714263e-05,
+      "loss": 1.4185,
+      "step": 12
+    },
+    {
+      "epoch": 0.04,
+      "grad_norm": 0.9233186885946617,
+      "learning_rate": 9.997747886043367e-05,
+      "loss": 1.3691,
+      "step": 13
+    },
+    {
+      "epoch": 0.04,
+      "grad_norm": 0.8728080646228755,
+      "learning_rate": 9.995996475642466e-05,
+      "loss": 1.3569,
+      "step": 14
+    },
+    {
+      "epoch": 0.05,
+      "grad_norm": 0.8307436189120798,
+      "learning_rate": 9.99374496282885e-05,
+      "loss": 1.3667,
+      "step": 15
+    },
+    {
+      "epoch": 0.05,
+      "grad_norm": 0.7939681765489442,
+      "learning_rate": 9.990993572980378e-05,
+      "loss": 1.3589,
+      "step": 16
+    },
+    {
+      "epoch": 0.05,
+      "grad_norm": 0.7980526587202514,
+      "learning_rate": 9.987742581512918e-05,
+      "loss": 1.395,
+      "step": 17
+    },
+    {
+      "epoch": 0.06,
+      "grad_norm": 0.7207026961586376,
+      "learning_rate": 9.983992313852774e-05,
+      "loss": 1.3887,
+      "step": 18
+    },
+    {
+      "epoch": 0.06,
+      "grad_norm": 0.7871310923443939,
+      "learning_rate": 9.979743145404119e-05,
+      "loss": 1.3062,
+      "step": 19
+    },
+    {
+      "epoch": 0.06,
+      "grad_norm": 0.7246807835643384,
+      "learning_rate": 9.974995501511404e-05,
+      "loss": 1.4028,
+      "step": 20
+    },
+    {
+      "epoch": 0.06,
+      "grad_norm": 0.7206361737972257,
+      "learning_rate": 9.969749857416789e-05,
+      "loss": 1.3398,
+      "step": 21
+    },
+    {
+      "epoch": 0.07,
+      "grad_norm": 0.6960105391155793,
+      "learning_rate": 9.964006738212575e-05,
+      "loss": 1.3469,
+      "step": 22
+    },
+    {
+      "epoch": 0.07,
+      "grad_norm": 0.7204455050781057,
+      "learning_rate": 9.957766718788633e-05,
+      "loss": 1.3765,
+      "step": 23
+    },
+    {
+      "epoch": 0.07,
+      "grad_norm": 0.5812596677574942,
+      "learning_rate": 9.951030423774859e-05,
+      "loss": 1.3379,
+      "step": 24
+    },
+    {
+      "epoch": 0.08,
+      "grad_norm": 0.675904894861765,
+      "learning_rate": 9.943798527478651e-05,
+      "loss": 1.3789,
+      "step": 25
+    },
+    {
+      "epoch": 0.08,
+      "grad_norm": 0.6664199931252878,
+      "learning_rate": 9.936071753817415e-05,
+      "loss": 1.3535,
+      "step": 26
+    },
+    {
+      "epoch": 0.08,
+      "grad_norm": 0.7030363325377104,
+      "learning_rate": 9.927850876246088e-05,
+      "loss": 1.218,
+      "step": 27
+    },
+    {
+      "epoch": 0.09,
+      "grad_norm": 0.6185910243439269,
+      "learning_rate": 9.919136717679722e-05,
+      "loss": 1.3501,
+      "step": 28
+    },
+    {
+      "epoch": 0.09,
+      "grad_norm": 0.6270496500828271,
+      "learning_rate": 9.909930150411113e-05,
+      "loss": 1.3628,
+      "step": 29
+    },
+    {
+      "epoch": 0.09,
+      "grad_norm": 0.6021577621428597,
+      "learning_rate": 9.900232096023477e-05,
+      "loss": 1.3428,
+      "step": 30
+    },
+    {
+      "epoch": 0.1,
+      "grad_norm": 0.6760900210464318,
+      "learning_rate": 9.890043525298203e-05,
+      "loss": 1.3079,
+      "step": 31
+    },
+    {
+      "epoch": 0.1,
+      "grad_norm": 0.5940959503401227,
+      "learning_rate": 9.879365458117678e-05,
+      "loss": 1.291,
+      "step": 32
+    },
+    {
+      "epoch": 0.1,
+      "grad_norm": 0.5427094594745479,
+      "learning_rate": 9.868198963363189e-05,
+      "loss": 1.2893,
+      "step": 33
+    },
+    {
+      "epoch": 0.1,
+      "grad_norm": 0.5958301777472343,
+      "learning_rate": 9.856545158807938e-05,
+      "loss": 1.2363,
+      "step": 34
+    },
+    {
+      "epoch": 0.11,
+      "grad_norm": 0.5850214722817281,
+      "learning_rate": 9.844405211005146e-05,
+      "loss": 1.3154,
+      "step": 35
+    },
+    {
+      "epoch": 0.11,
+      "grad_norm": 0.6172288827652985,
+      "learning_rate": 9.831780335171279e-05,
+      "loss": 1.3101,
+      "step": 36
+    },
+    {
+      "epoch": 0.11,
+      "grad_norm": 0.5728085013088785,
+      "learning_rate": 9.818671795064404e-05,
+      "loss": 1.2817,
+      "step": 37
+    },
+    {
+      "epoch": 0.12,
+      "grad_norm": 0.5979780148525169,
+      "learning_rate": 9.805080902857699e-05,
+      "loss": 1.2493,
+      "step": 38
+    },
+    {
+      "epoch": 0.12,
+      "grad_norm": 0.537951216091522,
+      "learning_rate": 9.791009019008078e-05,
+      "loss": 1.2454,
+      "step": 39
+    },
+    {
+      "epoch": 0.12,
+      "grad_norm": 0.5879388223555754,
+      "learning_rate": 9.776457552120033e-05,
+      "loss": 1.2744,
+      "step": 40
+    },
+    {
+      "epoch": 0.13,
+      "grad_norm": 0.5828144715679795,
+      "learning_rate": 9.761427958804621e-05,
+      "loss": 1.2354,
+      "step": 41
+    },
+    {
+      "epoch": 0.13,
+      "grad_norm": 0.5640545987289305,
+      "learning_rate": 9.745921743533653e-05,
+      "loss": 1.3105,
+      "step": 42
+    },
+    {
+      "epoch": 0.13,
+      "grad_norm": 0.6077402891626962,
+      "learning_rate": 9.729940458489104e-05,
+      "loss": 1.3188,
+      "step": 43
+    },
+    {
+      "epoch": 0.14,
+      "grad_norm": 0.5649165442230879,
+      "learning_rate": 9.713485703407731e-05,
+      "loss": 1.2756,
+      "step": 44
+    },
+    {
+      "epoch": 0.14,
+      "grad_norm": 0.5751090024359722,
+      "learning_rate": 9.696559125420948e-05,
+      "loss": 1.2336,
+      "step": 45
+    },
+    {
+      "epoch": 0.14,
+      "grad_norm": 0.5415492855376868,
+      "learning_rate": 9.679162418889931e-05,
+      "loss": 1.2424,
+      "step": 46
+    },
+    {
+      "epoch": 0.14,
+      "grad_norm": 0.5742691889803935,
+      "learning_rate": 9.66129732523603e-05,
+      "loss": 1.2363,
+      "step": 47
+    },
+    {
+      "epoch": 0.15,
+      "grad_norm": 0.5763001671560306,
+      "learning_rate": 9.642965632766436e-05,
+      "loss": 1.2661,
+      "step": 48
+    },
+    {
+      "epoch": 0.15,
+      "grad_norm": 0.5940077836695673,
+      "learning_rate": 9.624169176495184e-05,
+      "loss": 1.2734,
+      "step": 49
+    },
+    {
+      "epoch": 0.15,
+      "grad_norm": 0.6005566973905004,
+      "learning_rate": 9.604909837959455e-05,
+      "loss": 1.2078,
+      "step": 50
+    },
+    {
+      "epoch": 0.16,
+      "grad_norm": 0.5429670344508561,
+      "learning_rate": 9.585189545031238e-05,
+      "loss": 1.2246,
+      "step": 51
+    },
+    {
+      "epoch": 0.16,
+      "grad_norm": 0.5871060174846529,
+      "learning_rate": 9.565010271724352e-05,
+      "loss": 1.3501,
+      "step": 52
+    },
+    {
+      "epoch": 0.16,
+      "grad_norm": 0.5907025473048154,
+      "learning_rate": 9.54437403799684e-05,
+      "loss": 1.2729,
+      "step": 53
+    },
+    {
+      "epoch": 0.17,
+      "grad_norm": 0.6516780485358238,
+      "learning_rate": 9.523282909548773e-05,
+      "loss": 1.2881,
+      "step": 54
+    },
+    {
+      "epoch": 0.17,
+      "grad_norm": 0.5670548168352986,
+      "learning_rate": 9.50173899761547e-05,
+      "loss": 1.2056,
+      "step": 55
+    },
+    {
+      "epoch": 0.17,
+      "grad_norm": 0.5390303783246289,
+      "learning_rate": 9.47974445875617e-05,
+      "loss": 1.2144,
+      "step": 56
+    },
+    {
+      "epoch": 0.18,
+      "grad_norm": 0.5646076772019145,
+      "learning_rate": 9.457301494638147e-05,
+      "loss": 1.2212,
+      "step": 57
+    },
+    {
+      "epoch": 0.18,
+      "grad_norm": 0.5612167041328082,
+      "learning_rate": 9.434412351816328e-05,
+      "loss": 1.2717,
+      "step": 58
+    },
+    {
+      "epoch": 0.18,
+      "grad_norm": 0.5263144027788276,
+      "learning_rate": 9.411079321508414e-05,
+      "loss": 1.2112,
+      "step": 59
+    },
+    {
+      "epoch": 0.18,
+      "grad_norm": 0.5356079649939604,
+      "learning_rate": 9.387304739365523e-05,
+      "loss": 1.2234,
+      "step": 60
+    },
+    {
+      "epoch": 0.19,
+      "grad_norm": 0.5615087563747105,
+      "learning_rate": 9.36309098523839e-05,
+      "loss": 1.2996,
+      "step": 61
+    },
+    {
+      "epoch": 0.19,
+      "grad_norm": 0.5314023973665841,
+      "learning_rate": 9.338440482939146e-05,
+      "loss": 1.1797,
+      "step": 62
+    },
+    {
+      "epoch": 0.19,
+      "grad_norm": 0.5539257915730327,
+      "learning_rate": 9.31335569999869e-05,
+      "loss": 1.2395,
+      "step": 63
+    },
+    {
+      "epoch": 0.2,
+      "grad_norm": 0.5648098416860655,
+      "learning_rate": 9.287839147419686e-05,
+      "loss": 1.2156,
+      "step": 64
+    },
+    {
+      "epoch": 0.2,
+      "grad_norm": 0.5865707594533118,
+      "learning_rate": 9.261893379425218e-05,
+      "loss": 1.2683,
+      "step": 65
+    },
+    {
+      "epoch": 0.2,
+      "grad_norm": 0.5440049015077518,
+      "learning_rate": 9.2355209932031e-05,
+      "loss": 1.1853,
+      "step": 66
+    },
+    {
+      "epoch": 0.21,
+      "grad_norm": 0.5440912007358205,
+      "learning_rate": 9.208724628645902e-05,
+      "loss": 1.207,
+      "step": 67
+    },
+    {
+      "epoch": 0.21,
+      "grad_norm": 0.5587040510710671,
+      "learning_rate": 9.181506968086697e-05,
+      "loss": 1.1768,
+      "step": 68
+    },
+    {
+      "epoch": 0.21,
+      "grad_norm": 0.5192024047252568,
+      "learning_rate": 9.153870736030548e-05,
+      "loss": 1.2427,
+      "step": 69
+    },
+    {
+      "epoch": 0.22,
+      "grad_norm": 0.528230045301502,
+      "learning_rate": 9.125818698881798e-05,
+      "loss": 1.2168,
+      "step": 70
+    },
+    {
+      "epoch": 0.22,
+      "grad_norm": 0.5562754356167662,
+      "learning_rate": 9.097353664667138e-05,
+      "loss": 1.2793,
+      "step": 71
+    },
+    {
+      "epoch": 0.22,
+      "grad_norm": 0.5415713371356312,
+      "learning_rate": 9.068478482754532e-05,
+      "loss": 1.2334,
+      "step": 72
+    },
+    {
+      "epoch": 0.22,
+      "grad_norm": 0.5683526535348254,
+      "learning_rate": 9.03919604356798e-05,
+      "loss": 1.2534,
+      "step": 73
+    },
+    {
+      "epoch": 0.23,
+      "grad_norm": 0.5396091696393609,
+      "learning_rate": 9.0095092782982e-05,
+      "loss": 1.2722,
+      "step": 74
+    },
+    {
+      "epoch": 0.23,
+      "grad_norm": 0.513589724983156,
+      "learning_rate": 8.979421158609206e-05,
+      "loss": 1.2083,
+      "step": 75
+    },
+    {
+      "epoch": 0.23,
+      "grad_norm": 0.5376666654796496,
+      "learning_rate": 8.948934696340843e-05,
+      "loss": 1.1931,
+      "step": 76
+    },
+    {
+      "epoch": 0.24,
+      "grad_norm": 0.5368664737517322,
+      "learning_rate": 8.918052943207298e-05,
+      "loss": 1.2202,
+      "step": 77
+    },
+    {
+      "epoch": 0.24,
+      "grad_norm": 0.5263413266320822,
+      "learning_rate": 8.886778990491631e-05,
+      "loss": 1.1553,
+      "step": 78
+    },
+    {
+      "epoch": 0.24,
+      "grad_norm": 0.5604541523029579,
+      "learning_rate": 8.85511596873632e-05,
+      "loss": 1.2964,
+      "step": 79
+    },
+    {
+      "epoch": 0.25,
+      "grad_norm": 0.5427504321680312,
+      "learning_rate": 8.823067047429907e-05,
+      "loss": 1.2896,
+      "step": 80
+    },
+    {
+      "epoch": 0.25,
+      "grad_norm": 0.5474275345159585,
+      "learning_rate": 8.790635434689721e-05,
+      "loss": 1.2612,
+      "step": 81
+    },
+    {
+      "epoch": 0.25,
+      "grad_norm": 0.5269654905955425,
+      "learning_rate": 8.757824376940746e-05,
+      "loss": 1.1821,
+      "step": 82
+    },
+    {
+      "epoch": 0.26,
+      "grad_norm": 0.518380839571753,
+      "learning_rate": 8.724637158590652e-05,
+      "loss": 1.2393,
+      "step": 83
+    },
+    {
+      "epoch": 0.26,
+      "grad_norm": 0.5158889921158187,
+      "learning_rate": 8.691077101701024e-05,
+      "loss": 1.2754,
+      "step": 84
+    },
+    {
+      "epoch": 0.26,
+      "grad_norm": 0.4935757204910804,
+      "learning_rate": 8.65714756565482e-05,
+      "loss": 1.2173,
+      "step": 85
+    },
+    {
+      "epoch": 0.26,
+      "grad_norm": 0.5653934280740491,
+      "learning_rate": 8.622851946820095e-05,
+      "loss": 1.2253,
+      "step": 86
+    },
+    {
+      "epoch": 0.27,
+      "grad_norm": 0.5250362872573285,
+      "learning_rate": 8.588193678210026e-05,
+      "loss": 1.1707,
+      "step": 87
+    },
+    {
+      "epoch": 0.27,
+      "grad_norm": 0.5446362423647451,
+      "learning_rate": 8.553176229139261e-05,
+      "loss": 1.1589,
+      "step": 88
+    },
+    {
+      "epoch": 0.27,
+      "grad_norm": 0.5142920460404157,
+      "learning_rate": 8.517803104876639e-05,
+      "loss": 1.1948,
+      "step": 89
+    },
+    {
+      "epoch": 0.28,
+      "grad_norm": 0.5317755291367837,
+      "learning_rate": 8.482077846294308e-05,
+      "loss": 1.1812,
+      "step": 90
+    },
+    {
+      "epoch": 0.28,
+      "grad_norm": 0.5147982642702206,
+      "learning_rate": 8.446004029513294e-05,
+      "loss": 1.2266,
+      "step": 91
+    },
+    {
+      "epoch": 0.28,
+      "grad_norm": 0.5096116915413985,
+      "learning_rate": 8.409585265545509e-05,
+      "loss": 1.179,
+      "step": 92
+    },
+    {
+      "epoch": 0.29,
+      "grad_norm": 0.49910238964612424,
+      "learning_rate": 8.372825199932304e-05,
+      "loss": 1.2485,
+      "step": 93
+    },
+    {
+      "epoch": 0.29,
+      "grad_norm": 0.5507929652198581,
+      "learning_rate": 8.335727512379534e-05,
+      "loss": 1.2388,
+      "step": 94
+    },
+    {
+      "epoch": 0.29,
+      "grad_norm": 0.5063227228339493,
+      "learning_rate": 8.298295916389234e-05,
+      "loss": 1.2415,
+      "step": 95
+    },
+    {
+      "epoch": 0.3,
+      "grad_norm": 0.48774944558422,
+      "learning_rate": 8.260534158887876e-05,
+      "loss": 1.1301,
+      "step": 96
+    },
+    {
+      "epoch": 0.3,
+      "grad_norm": 0.5131731261481731,
+      "learning_rate": 8.222446019851314e-05,
+      "loss": 1.2158,
+      "step": 97
+    },
+    {
+      "epoch": 0.3,
+      "grad_norm": 0.49068550228239954,
+      "learning_rate": 8.184035311926396e-05,
+      "loss": 1.1968,
+      "step": 98
+    },
+    {
+      "epoch": 0.3,
+      "grad_norm": 0.5288695654598018,
+      "learning_rate": 8.145305880049328e-05,
+      "loss": 1.2637,
+      "step": 99
+    },
+    {
+      "epoch": 0.31,
+      "grad_norm": 0.5479388159051014,
+      "learning_rate": 8.106261601060772e-05,
+      "loss": 1.3218,
+      "step": 100
+    },
+    {
+      "epoch": 0.31,
+      "grad_norm": 0.5621161336370604,
+      "learning_rate": 8.066906383317801e-05,
+      "loss": 1.1729,
+      "step": 101
+    },
+    {
+      "epoch": 0.31,
+      "grad_norm": 0.4924846556483893,
+      "learning_rate": 8.027244166302642e-05,
+      "loss": 1.1875,
+      "step": 102
+    },
+    {
+      "epoch": 0.32,
+      "grad_norm": 0.5037791405351364,
+      "learning_rate": 7.987278920228349e-05,
+      "loss": 1.2539,
+      "step": 103
+    },
+    {
+      "epoch": 0.32,
+      "grad_norm": 0.522728960991421,
+      "learning_rate": 7.947014645641379e-05,
+      "loss": 1.2217,
+      "step": 104
+    },
+    {
+      "epoch": 0.32,
+      "grad_norm": 0.5182027559256257,
+      "learning_rate": 7.906455373021129e-05,
+      "loss": 1.2188,
+      "step": 105
+    },
+    {
+      "epoch": 0.33,
+      "grad_norm": 0.5539472403444413,
+      "learning_rate": 7.865605162376486e-05,
+      "loss": 1.1509,
+      "step": 106
+    },
+    {
+      "epoch": 0.33,
+      "grad_norm": 0.5439005321857965,
+      "learning_rate": 7.824468102839419e-05,
+      "loss": 1.251,
+      "step": 107
+    },
+    {
+      "epoch": 0.33,
+      "grad_norm": 0.5093593984665626,
+      "learning_rate": 7.783048312255653e-05,
+      "loss": 1.229,
+      "step": 108
+    },
+    {
+      "epoch": 0.34,
+      "grad_norm": 0.5145253165856254,
+      "learning_rate": 7.741349936772469e-05,
+      "loss": 1.1824,
+      "step": 109
+    },
+    {
+      "epoch": 0.34,
+      "grad_norm": 0.5702243863746732,
+      "learning_rate": 7.699377150423672e-05,
+      "loss": 1.1582,
+      "step": 110
+    },
+    {
+      "epoch": 0.34,
+      "grad_norm": 0.5451183103162889,
+      "learning_rate": 7.65713415471177e-05,
+      "loss": 1.2122,
+      "step": 111
+    },
+    {
+      "epoch": 0.34,
+      "grad_norm": 0.5252426501201418,
+      "learning_rate": 7.614625178187402e-05,
+      "loss": 1.1833,
+      "step": 112
+    },
+    {
+      "epoch": 0.35,
+      "grad_norm": 0.48749296422821287,
+      "learning_rate": 7.571854476026048e-05,
+      "loss": 1.2705,
+      "step": 113
+    },
+    {
+      "epoch": 0.35,
+      "grad_norm": 0.502002191255964,
+      "learning_rate": 7.528826329602099e-05,
+      "loss": 1.2188,
+      "step": 114
+    },
+    {
+      "epoch": 0.35,
+      "grad_norm": 0.48918503246260797,
+      "learning_rate": 7.485545046060271e-05,
+      "loss": 1.1997,
+      "step": 115
+    },
+    {
+      "epoch": 0.36,
+      "grad_norm": 0.5000600132443745,
+      "learning_rate": 7.442014957884472e-05,
+      "loss": 1.2051,
+      "step": 116
+    },
+    {
+      "epoch": 0.36,
+      "grad_norm": 0.5463372631998543,
+      "learning_rate": 7.398240422464109e-05,
+      "loss": 1.2214,
+      "step": 117
+    },
+    {
+      "epoch": 0.36,
+      "grad_norm": 0.4961344121724833,
+      "learning_rate": 7.354225821657914e-05,
+      "loss": 1.208,
+      "step": 118
+    },
+    {
+      "epoch": 0.37,
+      "grad_norm": 0.5899278757904003,
+      "learning_rate": 7.309975561355312e-05,
+      "loss": 1.146,
+      "step": 119
+    },
+    {
+      "epoch": 0.37,
+      "grad_norm": 0.5023010444635025,
+      "learning_rate": 7.265494071035401e-05,
+      "loss": 1.1509,
+      "step": 120
+    },
+    {
+      "epoch": 0.37,
+      "grad_norm": 0.5098342526132932,
+      "learning_rate": 7.220785803323544e-05,
+      "loss": 1.1743,
+      "step": 121
+    },
+    {
+      "epoch": 0.38,
+      "grad_norm": 0.5181510224338065,
+      "learning_rate": 7.175855233545668e-05,
+      "loss": 1.1807,
+      "step": 122
+    },
+    {
+      "epoch": 0.38,
+      "grad_norm": 0.5126009147492929,
+      "learning_rate": 7.130706859280274e-05,
+      "loss": 1.1875,
+      "step": 123
+    },
+    {
+      "epoch": 0.38,
+      "grad_norm": 0.5694155609516423,
+      "learning_rate": 7.085345199908235e-05,
+      "loss": 1.1428,
+      "step": 124
+    },
+    {
+      "epoch": 0.38,
+      "grad_norm": 0.5320176914945207,
+      "learning_rate": 7.03977479616039e-05,
+      "loss": 1.219,
+      "step": 125
+    },
+    {
+      "epoch": 0.39,
+      "grad_norm": 0.5263190337396886,
+      "learning_rate": 6.994000209663036e-05,
+      "loss": 1.1709,
+      "step": 126
+    },
+    {
+      "epoch": 0.39,
+      "grad_norm": 0.5043242059465939,
+      "learning_rate": 6.948026022481279e-05,
+      "loss": 1.1968,
+      "step": 127
+    },
+    {
+      "epoch": 0.39,
+      "grad_norm": 0.5580562389957314,
+      "learning_rate": 6.901856836660386e-05,
+      "loss": 1.1494,
+      "step": 128
+    },
+    {
+      "epoch": 0.4,
+      "grad_norm": 0.495782343949626,
+      "learning_rate": 6.855497273765112e-05,
+      "loss": 1.2119,
+      "step": 129
+    },
+    {
+      "epoch": 0.4,
+      "grad_norm": 0.49510379613102157,
+      "learning_rate": 6.808951974417078e-05,
+      "loss": 1.2063,
+      "step": 130
+    },
+    {
+      "epoch": 0.4,
+      "grad_norm": 0.5536322418977505,
+      "learning_rate": 6.762225597830237e-05,
+      "loss": 1.2617,
+      "step": 131
+    },
+    {
+      "epoch": 0.41,
+      "grad_norm": 0.5668665640341709,
+      "learning_rate": 6.715322821344494e-05,
+      "loss": 1.1443,
+      "step": 132
+    },
+    {
+      "epoch": 0.41,
+      "grad_norm": 0.5083973254687169,
+      "learning_rate": 6.668248339957491e-05,
+      "loss": 1.2588,
+      "step": 133
+    },
+    {
+      "epoch": 0.41,
+      "grad_norm": 0.5188609075674584,
+      "learning_rate": 6.621006865854644e-05,
+      "loss": 1.1726,
+      "step": 134
+    },
+    {
+      "epoch": 0.42,
+      "grad_norm": 0.5083154005353754,
+      "learning_rate": 6.573603127937442e-05,
+      "loss": 1.1953,
+      "step": 135
+    },
+    {
+      "epoch": 0.42,
+      "grad_norm": 0.544802137981313,
+      "learning_rate": 6.526041871350086e-05,
+      "loss": 1.1853,
+      "step": 136
+    },
+    {
+      "epoch": 0.42,
+      "grad_norm": 0.4692918589408204,
+      "learning_rate": 6.478327857004495e-05,
+      "loss": 1.1267,
+      "step": 137
+    },
+    {
+      "epoch": 0.42,
+      "grad_norm": 0.504019443454899,
+      "learning_rate": 6.43046586110374e-05,
+      "loss": 1.2085,
+      "step": 138
+    },
+    {
+      "epoch": 0.43,
+      "grad_norm": 0.5071221863228176,
+      "learning_rate": 6.382460674663932e-05,
+      "loss": 1.2026,
+      "step": 139
+    },
+    {
+      "epoch": 0.43,
+      "grad_norm": 0.5146798822767533,
+      "learning_rate": 6.334317103034652e-05,
+      "loss": 1.1177,
+      "step": 140
+    },
+    {
+      "epoch": 0.43,
+      "grad_norm": 0.4996775842843673,
+      "learning_rate": 6.286039965417925e-05,
+      "loss": 1.1711,
+      "step": 141
+    },
+    {
+      "epoch": 0.44,
+      "grad_norm": 0.5288823443570614,
+      "learning_rate": 6.237634094385813e-05,
+      "loss": 1.1199,
+      "step": 142
+    },
+    {
+      "epoch": 0.44,
+      "grad_norm": 0.5214584824734313,
+      "learning_rate": 6.18910433539668e-05,
+      "loss": 1.1199,
+      "step": 143
+    },
+    {
+      "epoch": 0.44,
+      "grad_norm": 0.5287174552858058,
+      "learning_rate": 6.140455546310148e-05,
+      "loss": 1.2134,
+      "step": 144
+    },
+    {
+      "epoch": 0.45,
+      "grad_norm": 0.5254574455344191,
+      "learning_rate": 6.0916925969008275e-05,
+      "loss": 1.2578,
+      "step": 145
+    },
+    {
+      "epoch": 0.45,
+      "grad_norm": 0.5189177716169284,
+      "learning_rate": 6.042820368370854e-05,
+      "loss": 1.2236,
+      "step": 146
+    },
+    {
+      "epoch": 0.45,
+      "grad_norm": 0.5680190447979175,
+      "learning_rate": 5.993843752861266e-05,
+      "loss": 1.1443,
+      "step": 147
+    },
+    {
+      "epoch": 0.46,
+      "grad_norm": 0.471841513545942,
+      "learning_rate": 5.944767652962309e-05,
+      "loss": 1.1733,
+      "step": 148
+    },
+    {
+      "epoch": 0.46,
+      "grad_norm": 0.5131643851734395,
+      "learning_rate": 5.895596981222678e-05,
+      "loss": 1.1592,
+      "step": 149
+    },
+    {
+      "epoch": 0.46,
+      "grad_norm": 0.5069414300839418,
+      "learning_rate": 5.8463366596577706e-05,
+      "loss": 1.2537,
+      "step": 150
+    },
+    {
+      "epoch": 0.46,
+      "grad_norm": 0.5239372199223642,
+      "learning_rate": 5.796991619256985e-05,
+      "loss": 1.1743,
+      "step": 151
+    },
+    {
+      "epoch": 0.47,
+      "grad_norm": 0.5387326952860074,
+      "learning_rate": 5.747566799490132e-05,
+      "loss": 1.2075,
+      "step": 152
+    },
+    {
+      "epoch": 0.47,
+      "grad_norm": 0.5224076525892895,
+      "learning_rate": 5.6980671478129853e-05,
+      "loss": 1.2515,
+      "step": 153
+    },
+    {
+      "epoch": 0.47,
+      "grad_norm": 0.5609147216869792,
+      "learning_rate": 5.648497619172042e-05,
+      "loss": 1.1836,
+      "step": 154
+    },
+    {
+      "epoch": 0.48,
+      "grad_norm": 0.5465488763290158,
+      "learning_rate": 5.5988631755085264e-05,
+      "loss": 1.1433,
+      "step": 155
+    },
+    {
+      "epoch": 0.48,
+      "grad_norm": 0.5079660248232113,
+      "learning_rate": 5.549168785261698e-05,
+      "loss": 1.1812,
+      "step": 156
+    },
+    {
+      "epoch": 0.48,
+      "grad_norm": 0.5717154092705009,
+      "learning_rate": 5.499419422871506e-05,
+      "loss": 1.2068,
+      "step": 157
+    },
+    {
+      "epoch": 0.49,
+      "grad_norm": 0.5301493442832448,
+      "learning_rate": 5.4496200682806495e-05,
+      "loss": 1.2273,
+      "step": 158
+    },
+    {
+      "epoch": 0.49,
+      "grad_norm": 0.5098890110687092,
+      "learning_rate": 5.399775706436076e-05,
+      "loss": 1.2134,
+      "step": 159
+    },
+    {
+      "epoch": 0.49,
+      "grad_norm": 0.5365690403371399,
+      "learning_rate": 5.3498913267899864e-05,
+      "loss": 1.2051,
+      "step": 160
+    },
+    {
+      "epoch": 0.5,
+      "grad_norm": 0.5002287628666306,
+      "learning_rate": 5.299971922800391e-05,
+      "loss": 1.1255,
+      "step": 161
+    },
+    {
+      "epoch": 0.5,
+      "grad_norm": 0.5276270904967929,
+      "learning_rate": 5.250022491431259e-05,
+      "loss": 1.2124,
+      "step": 162
+    },
+    {
+      "epoch": 0.5,
+      "grad_norm": 0.5224515486361312,
+      "learning_rate": 5.200048032652318e-05,
+      "loss": 1.2559,
+      "step": 163
+    },
+    {
+      "epoch": 0.5,
+      "grad_norm": 0.49316666866194575,
+      "learning_rate": 5.150053548938557e-05,
+      "loss": 1.1421,
+      "step": 164
+    },
+    {
+      "epoch": 0.51,
+      "grad_norm": 0.5388196993705506,
+      "learning_rate": 5.100044044769472e-05,
+      "loss": 1.2017,
+      "step": 165
+    },
+    {
+      "epoch": 0.51,
+      "grad_norm": 0.4944161877205336,
+      "learning_rate": 5.0500245261281175e-05,
+      "loss": 1.1838,
+      "step": 166
+    },
+    {
+      "epoch": 0.51,
+      "grad_norm": 0.5075949586376647,
+      "learning_rate": 5e-05,
+      "loss": 1.2698,
+      "step": 167
+    },
+    {
+      "epoch": 0.52,
+      "grad_norm": 0.5288349964384436,
+      "learning_rate": 4.949975473871884e-05,
+      "loss": 1.2051,
+      "step": 168
+    },
+    {
+      "epoch": 0.52,
+      "grad_norm": 0.5298564213314307,
+      "learning_rate": 4.899955955230529e-05,
+      "loss": 1.1921,
+      "step": 169
+    },
+    {
+      "epoch": 0.52,
+      "grad_norm": 0.5006832164442894,
+      "learning_rate": 4.849946451061443e-05,
+      "loss": 1.0999,
+      "step": 170
+    },
+    {
+      "epoch": 0.53,
+      "grad_norm": 0.5180365116427443,
+      "learning_rate": 4.799951967347683e-05,
+      "loss": 1.1248,
+      "step": 171
+    },
+    {
+      "epoch": 0.53,
+      "grad_norm": 0.5085456414296031,
+      "learning_rate": 4.749977508568742e-05,
+      "loss": 1.0989,
+      "step": 172
+    },
+    {
+      "epoch": 0.53,
+      "grad_norm": 0.5062569741436713,
+      "learning_rate": 4.7000280771996104e-05,
+      "loss": 1.1851,
+      "step": 173
+    },
+    {
+      "epoch": 0.54,
+      "grad_norm": 0.49915526983252295,
+      "learning_rate": 4.650108673210015e-05,
+      "loss": 1.1565,
+      "step": 174
+    },
+    {
+      "epoch": 0.54,
+      "grad_norm": 0.530737295373366,
+      "learning_rate": 4.6002242935639254e-05,
+      "loss": 1.1367,
+      "step": 175
+    },
+    {
+      "epoch": 0.54,
+      "grad_norm": 0.5354173274123866,
+      "learning_rate": 4.550379931719351e-05,
+      "loss": 1.2231,
+      "step": 176
+    },
+    {
+      "epoch": 0.54,
+      "grad_norm": 0.4910072525351625,
+      "learning_rate": 4.500580577128495e-05,
+      "loss": 1.2354,
+      "step": 177
+    },
+    {
+      "epoch": 0.55,
+      "grad_norm": 0.5227332856552718,
+      "learning_rate": 4.4508312147383036e-05,
+      "loss": 1.2178,
+      "step": 178
+    },
+    {
+      "epoch": 0.55,
+      "grad_norm": 0.5288368214749951,
+      "learning_rate": 4.4011368244914755e-05,
+      "loss": 1.1731,
+      "step": 179
+    },
+    {
+      "epoch": 0.55,
+      "grad_norm": 0.5044952842565356,
+      "learning_rate": 4.3515023808279586e-05,
+      "loss": 1.1287,
+      "step": 180
+    },
+    {
+      "epoch": 0.56,
+      "grad_norm": 0.5325342803705855,
+      "learning_rate": 4.301932852187016e-05,
+      "loss": 1.2129,
+      "step": 181
+    },
+    {
+      "epoch": 0.56,
+      "grad_norm": 0.5415040662560654,
+      "learning_rate": 4.252433200509869e-05,
+      "loss": 1.3086,
+      "step": 182
+    },
+    {
+      "epoch": 0.56,
+      "grad_norm": 0.51763854212227,
+      "learning_rate": 4.203008380743016e-05,
+      "loss": 1.1216,
+      "step": 183
+    },
+    {
+      "epoch": 0.57,
+      "grad_norm": 0.5422430582487087,
+      "learning_rate": 4.1536633403422306e-05,
+      "loss": 1.1111,
+      "step": 184
+    },
+    {
+      "epoch": 0.57,
+      "grad_norm": 0.5022269460427503,
+      "learning_rate": 4.104403018777323e-05,
+      "loss": 1.1963,
+      "step": 185
+    },
+    {
+      "epoch": 0.57,
+      "grad_norm": 0.5631689643987351,
+      "learning_rate": 4.0552323470376916e-05,
+      "loss": 1.261,
+      "step": 186
+    },
+    {
+      "epoch": 0.58,
+      "grad_norm": 0.5083595167881586,
+      "learning_rate": 4.006156247138736e-05,
+      "loss": 1.2529,
+      "step": 187
+    },
+    {
+      "epoch": 0.58,
+      "grad_norm": 0.4982995201328196,
+      "learning_rate": 3.9571796316291476e-05,
+      "loss": 1.1938,
+      "step": 188
+    },
+    {
+      "epoch": 0.58,
+      "grad_norm": 0.5142407248745989,
+      "learning_rate": 3.908307403099174e-05,
+      "loss": 1.231,
+      "step": 189
+    },
+    {
+      "epoch": 0.58,
+      "grad_norm": 0.49845909814685674,
+      "learning_rate": 3.859544453689853e-05,
+      "loss": 1.1174,
+      "step": 190
+    },
+    {
+      "epoch": 0.59,
+      "grad_norm": 0.5044393346608876,
+      "learning_rate": 3.810895664603321e-05,
+      "loss": 1.1626,
+      "step": 191
+    },
+    {
+      "epoch": 0.59,
+      "grad_norm": 0.5028906716759755,
+      "learning_rate": 3.762365905614187e-05,
+      "loss": 1.1248,
+      "step": 192
+    },
+    {
+      "epoch": 0.59,
+      "grad_norm": 0.5052094751619233,
+      "learning_rate": 3.713960034582077e-05,
+      "loss": 1.1782,
+      "step": 193
+    },
+    {
+      "epoch": 0.6,
+      "grad_norm": 0.4927933528160678,
+      "learning_rate": 3.665682896965349e-05,
+      "loss": 1.1411,
+      "step": 194
+    },
+    {
+      "epoch": 0.6,
+      "grad_norm": 0.5194837061389329,
+      "learning_rate": 3.61753932533607e-05,
+      "loss": 1.1194,
+      "step": 195
+    },
+    {
+      "epoch": 0.6,
+      "grad_norm": 0.5003507430931593,
+      "learning_rate": 3.5695341388962614e-05,
+      "loss": 1.1438,
+      "step": 196
+    },
+    {
+      "epoch": 0.61,
+      "grad_norm": 0.4952909851844929,
+      "learning_rate": 3.521672142995506e-05,
+      "loss": 1.1274,
+      "step": 197
+    },
+    {
+      "epoch": 0.61,
+      "grad_norm": 0.515058338420701,
+      "learning_rate": 3.473958128649915e-05,
+      "loss": 1.2173,
+      "step": 198
+    },
+    {
+      "epoch": 0.61,
+      "grad_norm": 0.4649441022469789,
+      "learning_rate": 3.4263968720625594e-05,
+      "loss": 1.0901,
+      "step": 199
+    },
+    {
+      "epoch": 0.62,
+      "grad_norm": 0.5257495392411118,
+      "learning_rate": 3.378993134145356e-05,
+      "loss": 1.2124,
+      "step": 200
+    },
+    {
+      "epoch": 0.62,
+      "grad_norm": 0.4932641665110047,
+      "learning_rate": 3.33175166004251e-05,
+      "loss": 1.2024,
+      "step": 201
+    },
+    {
+      "epoch": 0.62,
+      "grad_norm": 0.47986279797474973,
+      "learning_rate": 3.284677178655507e-05,
+      "loss": 1.199,
+      "step": 202
+    },
+    {
+      "epoch": 0.62,
+      "grad_norm": 0.5192526802216362,
+      "learning_rate": 3.2377744021697644e-05,
+      "loss": 1.1348,
+      "step": 203
+    },
+    {
+      "epoch": 0.63,
+      "grad_norm": 0.49075397683891514,
+      "learning_rate": 3.1910480255829237e-05,
+      "loss": 1.209,
+      "step": 204
+    },
+    {
+      "epoch": 0.63,
+      "grad_norm": 0.5308583566356849,
+      "learning_rate": 3.1445027262348894e-05,
+      "loss": 1.1843,
+      "step": 205
+    },
+    {
+      "epoch": 0.63,
+      "grad_norm": 0.4754800140640908,
+      "learning_rate": 3.098143163339615e-05,
+      "loss": 1.0713,
+      "step": 206
+    },
+    {
+      "epoch": 0.64,
+      "grad_norm": 0.48744339587851926,
+      "learning_rate": 3.051973977518723e-05,
+      "loss": 1.1589,
+      "step": 207
+    },
+    {
+      "epoch": 0.64,
+      "grad_norm": 0.538169296667757,
+      "learning_rate": 3.0059997903369656e-05,
+      "loss": 1.1636,
+      "step": 208
+    },
+    {
+      "epoch": 0.64,
+      "grad_norm": 0.5174008006678488,
+      "learning_rate": 2.9602252038396093e-05,
+      "loss": 1.0088,
+      "step": 209
+    },
+    {
+      "epoch": 0.65,
+      "grad_norm": 0.5423637270068683,
+      "learning_rate": 2.914654800091768e-05,
+      "loss": 1.2637,
+      "step": 210
+    },
+    {
+      "epoch": 0.65,
+      "grad_norm": 0.49532340168296196,
+      "learning_rate": 2.8692931407197275e-05,
+      "loss": 1.1211,
+      "step": 211
+    },
+    {
+      "epoch": 0.65,
+      "grad_norm": 0.5223772948936137,
+      "learning_rate": 2.824144766454333e-05,
+      "loss": 1.1392,
+      "step": 212
+    },
+    {
+      "epoch": 0.66,
+      "grad_norm": 0.5301272448221841,
+      "learning_rate": 2.7792141966764568e-05,
+      "loss": 1.1663,
+      "step": 213
+    },
+    {
+      "epoch": 0.66,
+      "grad_norm": 0.5241453393589025,
+      "learning_rate": 2.7345059289646008e-05,
+      "loss": 1.1375,
+      "step": 214
+    },
+    {
+      "epoch": 0.66,
+      "grad_norm": 0.4946190439071369,
+      "learning_rate": 2.6900244386446904e-05,
+      "loss": 1.1843,
+      "step": 215
+    },
+    {
+      "epoch": 0.66,
+      "grad_norm": 0.4894503682723616,
+      "learning_rate": 2.6457741783420886e-05,
+      "loss": 1.1492,
+      "step": 216
+    },
+    {
+      "epoch": 0.67,
+      "grad_norm": 0.4852212167430057,
+      "learning_rate": 2.6017595775358926e-05,
+      "loss": 1.1592,
+      "step": 217
+    },
+    {
+      "epoch": 0.67,
+      "grad_norm": 0.5560815990000127,
+      "learning_rate": 2.5579850421155293e-05,
+      "loss": 1.1484,
+      "step": 218
+    },
+    {
+      "epoch": 0.67,
+      "grad_norm": 0.5035735515547093,
+      "learning_rate": 2.514454953939731e-05,
+      "loss": 1.1381,
+      "step": 219
+    },
+    {
+      "epoch": 0.68,
+      "grad_norm": 0.5057189594248047,
+      "learning_rate": 2.4711736703979018e-05,
+      "loss": 1.2036,
+      "step": 220
+    },
+    {
+      "epoch": 0.68,
+      "grad_norm": 0.5116012961964514,
+      "learning_rate": 2.428145523973952e-05,
+      "loss": 1.2212,
+      "step": 221
+    },
+    {
+      "epoch": 0.68,
+      "grad_norm": 0.48593835956939924,
+      "learning_rate": 2.3853748218126e-05,
+      "loss": 1.1505,
+      "step": 222
+    },
+    {
+      "epoch": 0.69,
+      "grad_norm": 0.5014137656412536,
+      "learning_rate": 2.342865845288232e-05,
+      "loss": 1.2246,
+      "step": 223
+    },
+    {
+      "epoch": 0.69,
+      "grad_norm": 0.5223070185696177,
+      "learning_rate": 2.3006228495763295e-05,
+      "loss": 1.158,
+      "step": 224
+    },
+    {
+      "epoch": 0.69,
+      "grad_norm": 0.5624122056478932,
+      "learning_rate": 2.258650063227533e-05,
+      "loss": 1.0566,
+      "step": 225
+    },
+    {
+      "epoch": 0.7,
+      "grad_norm": 0.5471967527021497,
+      "learning_rate": 2.2169516877443485e-05,
+      "loss": 1.1277,
+      "step": 226
+    },
+    {
+      "epoch": 0.7,
+      "grad_norm": 0.48337168913421796,
+      "learning_rate": 2.1755318971605826e-05,
+      "loss": 1.1956,
+      "step": 227
+    },
+    {
+      "epoch": 0.7,
+      "grad_norm": 0.5284626857313452,
+      "learning_rate": 2.1343948376235144e-05,
+      "loss": 1.196,
+      "step": 228
+    },
+    {
+      "epoch": 0.71,
+      "grad_norm": 0.5837213719579207,
+      "learning_rate": 2.0935446269788717e-05,
+      "loss": 1.0967,
+      "step": 229
+    },
+    {
+      "epoch": 0.71,
+      "grad_norm": 0.5165785735037434,
+      "learning_rate": 2.052985354358622e-05,
+      "loss": 1.2036,
+      "step": 230
+    },
+    {
+      "epoch": 0.71,
+      "grad_norm": 0.5536196580249753,
+      "learning_rate": 2.0127210797716524e-05,
+      "loss": 1.1572,
+      "step": 231
+    },
+    {
+      "epoch": 0.71,
+      "grad_norm": 0.4991948055619459,
+      "learning_rate": 1.9727558336973595e-05,
+      "loss": 1.2112,
+      "step": 232
+    },
+    {
+      "epoch": 0.72,
+      "grad_norm": 0.5271303549814744,
+      "learning_rate": 1.933093616682201e-05,
+      "loss": 1.0928,
+      "step": 233
+    },
+    {
+      "epoch": 0.72,
+      "grad_norm": 0.4755907436318383,
+      "learning_rate": 1.8937383989392294e-05,
+      "loss": 1.198,
+      "step": 234
+    },
+    {
+      "epoch": 0.72,
+      "grad_norm": 0.5071149662683123,
+      "learning_rate": 1.854694119950675e-05,
+      "loss": 1.1602,
+      "step": 235
+    },
+    {
+      "epoch": 0.73,
+      "grad_norm": 0.541369836697391,
+      "learning_rate": 1.8159646880736036e-05,
+      "loss": 1.1282,
+      "step": 236
+    },
+    {
+      "epoch": 0.73,
+      "grad_norm": 0.5083789491431415,
+      "learning_rate": 1.7775539801486867e-05,
+      "loss": 1.0872,
+      "step": 237
+    },
+    {
+      "epoch": 0.73,
+      "grad_norm": 0.5324743232408313,
+      "learning_rate": 1.739465841112125e-05,
+      "loss": 1.2222,
+      "step": 238
+    },
+    {
+      "epoch": 0.74,
+      "grad_norm": 0.48387142611649014,
+      "learning_rate": 1.701704083610768e-05,
+      "loss": 1.158,
+      "step": 239
+    },
+    {
+      "epoch": 0.74,
+      "grad_norm": 0.5141063576503123,
+      "learning_rate": 1.664272487620466e-05,
+      "loss": 1.1162,
+      "step": 240
+    },
+    {
+      "epoch": 0.74,
+      "grad_norm": 0.5408323267729829,
+      "learning_rate": 1.6271748000676983e-05,
+      "loss": 1.02,
+      "step": 241
+    },
+    {
+      "epoch": 0.75,
+      "grad_norm": 0.49627165191053163,
+      "learning_rate": 1.590414734454493e-05,
+      "loss": 1.0848,
+      "step": 242
+    },
+    {
+      "epoch": 0.75,
+      "grad_norm": 0.5376377474797022,
+      "learning_rate": 1.5539959704867085e-05,
+      "loss": 1.2019,
+      "step": 243
+    },
+    {
+      "epoch": 0.75,
+      "grad_norm": 0.49371506434647033,
+      "learning_rate": 1.517922153705692e-05,
+      "loss": 1.2151,
+      "step": 244
+    },
+    {
+      "epoch": 0.75,
+      "grad_norm": 0.523592301439241,
+      "learning_rate": 1.4821968951233638e-05,
+      "loss": 1.231,
+      "step": 245
+    },
+    {
+      "epoch": 0.76,
+      "grad_norm": 0.5074105674850246,
+      "learning_rate": 1.4468237708607397e-05,
+      "loss": 1.1519,
+      "step": 246
+    },
+    {
+      "epoch": 0.76,
+      "grad_norm": 0.5066768606133761,
+      "learning_rate": 1.4118063217899746e-05,
+      "loss": 1.1492,
+      "step": 247
+    },
+    {
+      "epoch": 0.76,
+      "grad_norm": 0.5116682098711993,
+      "learning_rate": 1.3771480531799052e-05,
+      "loss": 1.1477,
+      "step": 248
+    },
+    {
+      "epoch": 0.77,
+      "grad_norm": 0.5352752339451914,
+      "learning_rate": 1.342852434345181e-05,
+      "loss": 1.0964,
+      "step": 249
+    },
+    {
+      "epoch": 0.77,
+      "grad_norm": 0.5418770700535751,
+      "learning_rate": 1.308922898298977e-05,
+      "loss": 1.0803,
+      "step": 250
+    },
+    {
+      "epoch": 0.77,
+      "grad_norm": 0.5157514248366463,
+      "learning_rate": 1.2753628414093489e-05,
+      "loss": 1.1614,
+      "step": 251
+    },
+    {
+      "epoch": 0.78,
+      "grad_norm": 0.5351720244073139,
+      "learning_rate": 1.2421756230592534e-05,
+      "loss": 1.2224,
+      "step": 252
+    },
+    {
+      "epoch": 0.78,
+      "grad_norm": 0.5214450914678166,
+      "learning_rate": 1.2093645653102786e-05,
+      "loss": 1.2302,
+      "step": 253
+    },
+    {
+      "epoch": 0.78,
+      "grad_norm": 0.5409585268429792,
+      "learning_rate": 1.1769329525700935e-05,
+      "loss": 1.2104,
+      "step": 254
+    },
+    {
+      "epoch": 0.79,
+      "grad_norm": 0.5220300809084372,
+      "learning_rate": 1.144884031263681e-05,
+      "loss": 1.1047,
+      "step": 255
+    },
+    {
+      "epoch": 0.79,
+      "grad_norm": 0.5376568328092315,
+      "learning_rate": 1.1132210095083694e-05,
+      "loss": 1.2002,
+      "step": 256
+    },
+    {
+      "epoch": 0.79,
+      "grad_norm": 0.5414025185087213,
+      "learning_rate": 1.081947056792702e-05,
+      "loss": 1.1833,
+      "step": 257
+    },
+    {
+      "epoch": 0.79,
+      "grad_norm": 0.49383703911170834,
+      "learning_rate": 1.0510653036591583e-05,
+      "loss": 1.199,
+      "step": 258
+    },
+    {
+      "epoch": 0.8,
+      "grad_norm": 0.4916712820340113,
+      "learning_rate": 1.0205788413907952e-05,
+      "loss": 1.1292,
+      "step": 259
+    },
+    {
+      "epoch": 0.8,
+      "grad_norm": 0.5168351602558782,
+      "learning_rate": 9.904907217018e-06,
+      "loss": 1.2712,
+      "step": 260
+    },
+    {
+      "epoch": 0.8,
+      "grad_norm": 0.4856958512300068,
+      "learning_rate": 9.608039564320209e-06,
+      "loss": 1.1277,
+      "step": 261
+    },
+    {
+      "epoch": 0.81,
+      "grad_norm": 0.5106017400451819,
+      "learning_rate": 9.31521517245469e-06,
+      "loss": 1.1108,
+      "step": 262
+    },
+    {
+      "epoch": 0.81,
+      "grad_norm": 0.5036900746833531,
+      "learning_rate": 9.026463353328613e-06,
+      "loss": 1.1748,
+      "step": 263
+    },
+    {
+      "epoch": 0.81,
+      "grad_norm": 0.5033379794982934,
+      "learning_rate": 8.741813011182014e-06,
+      "loss": 1.1521,
+      "step": 264
+    },
+    {
+      "epoch": 0.82,
+      "grad_norm": 0.5332335505700774,
+      "learning_rate": 8.461292639694518e-06,
+      "loss": 1.0955,
+      "step": 265
+    },
+    {
+      "epoch": 0.82,
+      "grad_norm": 0.5125213937729919,
+      "learning_rate": 8.18493031913305e-06,
+      "loss": 1.1089,
+      "step": 266
+    },
+    {
+      "epoch": 0.82,
+      "grad_norm": 0.5148264028992737,
+      "learning_rate": 7.912753713540988e-06,
+      "loss": 1.177,
+      "step": 267
+    },
+    {
+      "epoch": 0.83,
+      "grad_norm": 0.5229543225675894,
+      "learning_rate": 7.644790067969005e-06,
+      "loss": 1.1646,
+      "step": 268
+    },
+    {
+      "epoch": 0.83,
+      "grad_norm": 0.541964343800654,
+      "learning_rate": 7.381066205747822e-06,
+      "loss": 1.1187,
+      "step": 269
+    },
+    {
+      "epoch": 0.83,
+      "grad_norm": 0.5266917236869202,
+      "learning_rate": 7.1216085258031414e-06,
+      "loss": 1.0901,
+      "step": 270
+    },
+    {
+      "epoch": 0.83,
+      "grad_norm": 0.49725116144459375,
+      "learning_rate": 6.866443000013117e-06,
+      "loss": 1.1526,
+      "step": 271
+    },
+    {
+      "epoch": 0.84,
+      "grad_norm": 0.5012370969649222,
+      "learning_rate": 6.6155951706085405e-06,
+      "loss": 1.1689,
+      "step": 272
+    },
+    {
+      "epoch": 0.84,
+      "grad_norm": 0.5129519814796674,
+      "learning_rate": 6.369090147616103e-06,
+      "loss": 1.1348,
+      "step": 273
+    },
+    {
+      "epoch": 0.84,
+      "grad_norm": 0.4926628244202811,
+      "learning_rate": 6.1269526063447765e-06,
+      "loss": 1.1538,
+      "step": 274
+    },
+    {
+      "epoch": 0.85,
+      "grad_norm": 0.5420926972187566,
+      "learning_rate": 5.889206784915862e-06,
+      "loss": 1.1484,
+      "step": 275
+    },
+    {
+      "epoch": 0.85,
+      "grad_norm": 0.4950562048422779,
+      "learning_rate": 5.6558764818367195e-06,
+      "loss": 1.2126,
+      "step": 276
+    },
+    {
+      "epoch": 0.85,
+      "grad_norm": 0.536783110803186,
+      "learning_rate": 5.426985053618544e-06,
+      "loss": 1.1301,
+      "step": 277
+    },
+    {
+      "epoch": 0.86,
+      "grad_norm": 0.5016583572718184,
+      "learning_rate": 5.2025554124383095e-06,
+      "loss": 1.1582,
+      "step": 278
+    },
+    {
+      "epoch": 0.86,
+      "grad_norm": 0.5263293678305013,
+      "learning_rate": 4.9826100238453135e-06,
+      "loss": 1.1577,
+      "step": 279
+    },
+    {
+      "epoch": 0.86,
+      "grad_norm": 0.5272007649322749,
+      "learning_rate": 4.767170904512292e-06,
+      "loss": 1.167,
+      "step": 280
+    },
+    {
+      "epoch": 0.87,
+      "grad_norm": 0.5063050835599103,
+      "learning_rate": 4.556259620031617e-06,
+      "loss": 1.1846,
+      "step": 281
+    },
+    {
+      "epoch": 0.87,
+      "grad_norm": 0.4910298010180289,
+      "learning_rate": 4.349897282756487e-06,
+      "loss": 1.1357,
+      "step": 282
+    },
+    {
+      "epoch": 0.87,
+      "grad_norm": 0.5379014923521229,
+      "learning_rate": 4.148104549687626e-06,
+      "loss": 1.0798,
+      "step": 283
+    },
+    {
+      "epoch": 0.87,
+      "grad_norm": 0.4865878079832316,
+      "learning_rate": 3.95090162040545e-06,
+      "loss": 1.1123,
+      "step": 284
+    },
+    {
+      "epoch": 0.88,
+      "grad_norm": 0.5282174531059731,
+      "learning_rate": 3.758308235048158e-06,
+      "loss": 1.1538,
+      "step": 285
+    },
+    {
+      "epoch": 0.88,
+      "grad_norm": 0.5010967709298223,
+      "learning_rate": 3.570343672335641e-06,
+      "loss": 1.1392,
+      "step": 286
+    },
+    {
+      "epoch": 0.88,
+      "grad_norm": 0.5189711142846094,
+      "learning_rate": 3.38702674763971e-06,
+      "loss": 1.1851,
+      "step": 287
+    },
+    {
+      "epoch": 0.89,
+      "grad_norm": 0.49664252429083205,
+      "learning_rate": 3.2083758111006945e-06,
+      "loss": 1.11,
+      "step": 288
+    },
+    {
+      "epoch": 0.89,
+      "grad_norm": 0.4859267407564821,
+      "learning_rate": 3.0344087457905346e-06,
+      "loss": 1.1719,
+      "step": 289
+    },
+    {
+      "epoch": 0.89,
+      "grad_norm": 0.5151543902141822,
+      "learning_rate": 2.86514296592269e-06,
+      "loss": 1.2102,
+      "step": 290
+    },
+    {
+      "epoch": 0.9,
+      "grad_norm": 0.5256996348656279,
+      "learning_rate": 2.7005954151089695e-06,
+      "loss": 1.1709,
+      "step": 291
+    },
+    {
+      "epoch": 0.9,
+      "grad_norm": 0.5102631445476464,
+      "learning_rate": 2.54078256466348e-06,
+      "loss": 1.2034,
+      "step": 292
+    },
+    {
+      "epoch": 0.9,
+      "grad_norm": 0.5093343267813872,
+      "learning_rate": 2.3857204119538014e-06,
+      "loss": 1.0969,
+      "step": 293
+    },
+    {
+      "epoch": 0.91,
+      "grad_norm": 0.5024898891939928,
+      "learning_rate": 2.2354244787996748e-06,
+      "loss": 1.1289,
+      "step": 294
+    },
+    {
+      "epoch": 0.91,
+      "grad_norm": 0.5279824056783474,
+      "learning_rate": 2.0899098099192273e-06,
+      "loss": 1.1672,
+      "step": 295
+    },
+    {
+      "epoch": 0.91,
+      "grad_norm": 0.5379712149279697,
+      "learning_rate": 1.9491909714230204e-06,
+      "loss": 1.1858,
+      "step": 296
+    },
+    {
+      "epoch": 0.91,
+      "grad_norm": 0.5436100119596193,
+      "learning_rate": 1.8132820493559521e-06,
+      "loss": 1.1201,
+      "step": 297
+    },
+    {
+      "epoch": 0.92,
+      "grad_norm": 0.5015550141769547,
+      "learning_rate": 1.6821966482872264e-06,
+      "loss": 1.1797,
+      "step": 298
+    },
+    {
+      "epoch": 0.92,
+      "grad_norm": 0.5119620808666944,
+      "learning_rate": 1.5559478899485447e-06,
+      "loss": 1.1493,
+      "step": 299
+    },
+    {
+      "epoch": 0.92,
+      "grad_norm": 0.4977128949907143,
+      "learning_rate": 1.434548411920622e-06,
+      "loss": 1.1821,
+      "step": 300
+    },
+    {
+      "epoch": 0.93,
+      "grad_norm": 0.509981327731476,
+      "learning_rate": 1.3180103663681165e-06,
+      "loss": 1.1377,
+      "step": 301
+    },
+    {
+      "epoch": 0.93,
+      "grad_norm": 0.5029662223988268,
+      "learning_rate": 1.206345418823235e-06,
+      "loss": 1.1526,
+      "step": 302
+    },
+    {
+      "epoch": 0.93,
+      "grad_norm": 0.5519080967379805,
+      "learning_rate": 1.099564747017967e-06,
+      "loss": 1.1282,
+      "step": 303
+    },
+    {
+      "epoch": 0.94,
+      "grad_norm": 0.5223711089586693,
+      "learning_rate": 9.976790397652315e-07,
+      "loss": 1.0811,
+      "step": 304
+    },
+    {
+      "epoch": 0.94,
+      "grad_norm": 0.47043707607640073,
+      "learning_rate": 9.006984958888742e-07,
+      "loss": 1.1685,
+      "step": 305
+    },
+    {
+      "epoch": 0.94,
+      "grad_norm": 0.5254608837175258,
+      "learning_rate": 8.086328232027873e-07,
+      "loss": 1.0779,
+      "step": 306
+    },
+    {
+      "epoch": 0.95,
+      "grad_norm": 0.5081931596861774,
+      "learning_rate": 7.214912375391291e-07,
+      "loss": 1.1371,
+      "step": 307
+    },
+    {
+      "epoch": 0.95,
+      "grad_norm": 0.4922428385525549,
+      "learning_rate": 6.392824618258519e-07,
+      "loss": 1.2107,
+      "step": 308
+    },
+    {
+      "epoch": 0.95,
+      "grad_norm": 0.5274751815037492,
+      "learning_rate": 5.620147252134889e-07,
+      "loss": 1.1548,
+      "step": 309
+    },
+    {
+      "epoch": 0.95,
+      "grad_norm": 0.4984402486182386,
+      "learning_rate": 4.896957622514298e-07,
+      "loss": 1.1189,
+      "step": 310
+    },
+    {
+      "epoch": 0.96,
+      "grad_norm": 0.503914938251447,
+      "learning_rate": 4.2233281211368493e-07,
+      "loss": 1.1362,
+      "step": 311
+    },
+    {
+      "epoch": 0.96,
+      "grad_norm": 0.5313082041540377,
+      "learning_rate": 3.599326178742535e-07,
+      "loss": 1.2146,
+      "step": 312
+    },
+    {
+      "epoch": 0.96,
+      "grad_norm": 0.5301442913189116,
+      "learning_rate": 3.025014258321135e-07,
+      "loss": 1.1699,
+      "step": 313
+    },
+    {
+      "epoch": 0.97,
+      "grad_norm": 0.5176192540816547,
+      "learning_rate": 2.500449848859776e-07,
+      "loss": 1.2048,
+      "step": 314
+    },
+    {
+      "epoch": 0.97,
+      "grad_norm": 0.5034214543341652,
+      "learning_rate": 2.0256854595881447e-07,
+      "loss": 1.1338,
+      "step": 315
+    },
+    {
+      "epoch": 0.97,
+      "grad_norm": 0.4795708437379455,
+      "learning_rate": 1.6007686147225254e-07,
+      "loss": 1.1533,
+      "step": 316
+    },
+    {
+      "epoch": 0.98,
+      "grad_norm": 0.5053834258977199,
+      "learning_rate": 1.2257418487082727e-07,
+      "loss": 1.1648,
+      "step": 317
+    },
+    {
+      "epoch": 0.98,
+      "grad_norm": 0.5276919175530018,
+      "learning_rate": 9.006427019622176e-08,
+      "loss": 1.2034,
+      "step": 318
+    },
+    {
+      "epoch": 0.98,
+      "grad_norm": 0.5495649945557277,
+      "learning_rate": 6.255037171150612e-08,
+      "loss": 1.1985,
+      "step": 319
+    },
+    {
+      "epoch": 0.99,
+      "grad_norm": 0.4946180370647359,
+      "learning_rate": 4.0035243575342605e-08,
+      "loss": 1.1738,
+      "step": 320
+    },
+    {
+      "epoch": 0.99,
+      "grad_norm": 0.47883695876558374,
+      "learning_rate": 2.2521139566328285e-08,
+      "loss": 1.1804,
+      "step": 321
+    },
+    {
+      "epoch": 0.99,
+      "grad_norm": 0.5584783758937163,
+      "learning_rate": 1.0009812857370016e-08,
+      "loss": 1.1968,
+      "step": 322
+    },
+    {
+      "epoch": 0.99,
+      "grad_norm": 0.4815115437532202,
+      "learning_rate": 2.502515840197006e-09,
+      "loss": 1.1697,
+      "step": 323
+    },
+    {
+      "epoch": 1.0,
+      "grad_norm": 0.5515369695100554,
+      "learning_rate": 0.0,
+      "loss": 1.1912,
+      "step": 324
+    },
+    {
+      "epoch": 1.0,
+      "step": 324,
+      "total_flos": 2.9498071531598643e+18,
+      "train_loss": 1.2066333912037037,
+      "train_runtime": 17698.3629,
+      "train_samples_per_second": 4.697,
+      "train_steps_per_second": 0.018
+    }
+  ],
+  "logging_steps": 1.0,
+  "max_steps": 324,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 1,
+  "save_steps": 20,
+  "total_flos": 2.9498071531598643e+18,
+  "train_batch_size": 16,
+  "trial_name": null,
+  "trial_params": null
+}