2

checkpoint-160/added_tokens.json

@@ -0,0 +1,3 @@
+{
+  "<image>": 32000
+}

{llavav16bta_cp3d1688_longb_qafull_llhound_fullimg_multiimg_trainllm_e1_bz512_32f1688 → checkpoint-160}/model-00001-of-00003.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:badbca0f8b6f276fcf9f0b6f35ad29370715f5d9e5bde69804ab78457e611ced
+oid sha256:74c7f5d993a912aa5407d99fe5b88b36c727a99eab963941297cac08432831c3
 size 4980641554

{llavav16bta_cp3d1688_longb_qafull_llhound_fullimg_trainllm_e1_bz256_32f1688 → checkpoint-160}/model-00002-of-00003.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:6093c7766f8f45b2f51be4ebb77ed2df5c49b410970e0620a6d5d4ff7e8b0c13
+oid sha256:0c2bdfeb77728fd9619f3bb3f4f7f5f4352ff82f3f730a314d4bd9d9dd641651
 size 4957878536

{llavav16bta_cp3d1688_longb_qafull_llhound_fullimg_trainllm_e1_bz256_32f1688 → checkpoint-160}/model-00003-of-00003.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e74b0d8a25e3c65d3965fcfdaf83dd54ef2f8087c14d12f202a80af25e0ad167
+oid sha256:b57c8234e6548525e69218c26509caed76ebf9911b1190db753fa142814aaa63
 size 4580905400

checkpoint-160/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
+  }
+}

llavav16bta_cp3d1688_longb_qafull_llhound_fullimg_multiimg_trainllm_e1_bz512_32f1688/training_args.bin → checkpoint-160/tokenizer.model

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:6c61e1b93e453464c82a151998343e20e07c872d423dcc4a4a2e46bd6e53948b
-size 7096
+oid sha256:9e556afd44213b6bd1be2b850ebbbd98f5481437a8021afaf58ee7fb1818d347
+size 499723

checkpoint-160/tokenizer_config.json

@@ -0,0 +1,52 @@
+{
+  "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
+    },
+    "32000": {
+      "content": "<image>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "bos_token": "<s>",
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "</s>",
+  "legacy": false,
+  "model_max_length": 2048,
+  "pad_token": "<unk>",
+  "padding_side": "right",
+  "processor_class": "LlavaNextProcessor",
+  "sp_model_kwargs": {},
+  "spaces_between_special_tokens": false,
+  "tokenizer_class": "LlamaTokenizer",
+  "unk_token": "<unk>",
+  "use_default_system_prompt": false
+}

{llavav16bta_cp3d1688_longb_qafull_llhound_fullimg_trainllm_e1_bz256_32f1688 → checkpoint-160}/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:3db21088a03dab3e9042367b9f7e3bb67dfc608a5563e28fb3fb76fa7ab52d6c
-size 7032
+oid sha256:6d53e83224f2769f2c964ded43fbc4dbe376aaf199dca61db75f92ac145c4ee2
+size 6712

llavav16bta_cp3d1688_longb_qafull_llhound_fullimg_multiimg_trainllm_e1_bz512_32f1688/model-00002-of-00003.safetensors

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:c486b79a3c76b9a43cfd068b592bfc88cfac2e5564d2552c55b24b518581eb0d
-size 4957878536

llavav16bta_cp3d1688_longb_qafull_llhound_fullimg_multiimg_trainllm_e1_bz512_32f1688/model-00003-of-00003.safetensors

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:e4df79bed0e8f9f55e5457c7bc44022a71d5a4d7f5e4402b4da25a356401f2cf
-size 4580905400

llavav16bta_cp3d1688_longb_qafull_llhound_fullimg_multiimg_trainllm_e1_bz512_32f1688/trainer_state.json

@@ -1,422 +0,0 @@
-{
-  "best_metric": null,
-  "best_model_checkpoint": null,
-  "epoch": 0.9999115122555526,
-  "eval_steps": 500,
-  "global_step": 2825,
-  "is_hyper_param_search": false,
-  "is_local_process_zero": true,
-  "is_world_process_zero": true,
-  "log_history": [
-    {
-      "epoch": 0.017697548889478807,
-      "grad_norm": 1.3246452072458292,
-      "learning_rate": 1.1764705882352942e-05,
-      "loss": 1.3482,
-      "step": 50
-    },
-    {
-      "epoch": 0.035395097778957614,
-      "grad_norm": 1.184986868355016,
-      "learning_rate": 1.9998521094455198e-05,
-      "loss": 1.0655,
-      "step": 100
-    },
-    {
-      "epoch": 0.053092646668436425,
-      "grad_norm": 0.9956602761532123,
-      "learning_rate": 1.9972241607451552e-05,
-      "loss": 1.0325,
-      "step": 150
-    },
-    {
-      "epoch": 0.07079019555791523,
-      "grad_norm": 0.8985264278021503,
-      "learning_rate": 1.9913196946839304e-05,
-      "loss": 1.0115,
-      "step": 200
-    },
-    {
-      "epoch": 0.08848774444739403,
-      "grad_norm": 0.9761536009178263,
-      "learning_rate": 1.9821581111985072e-05,
-      "loss": 1.0087,
-      "step": 250
-    },
-    {
-      "epoch": 0.10618529333687285,
-      "grad_norm": 0.9582085355410745,
-      "learning_rate": 1.9697695119335547e-05,
-      "loss": 0.9959,
-      "step": 300
-    },
-    {
-      "epoch": 0.12388284222635165,
-      "grad_norm": 0.9502507945345694,
-      "learning_rate": 1.954194601338651e-05,
-      "loss": 0.9832,
-      "step": 350
-    },
-    {
-      "epoch": 0.14158039111583046,
-      "grad_norm": 0.9312008953589136,
-      "learning_rate": 1.9354845529282042e-05,
-      "loss": 0.9838,
-      "step": 400
-    },
-    {
-      "epoch": 0.15927794000530926,
-      "grad_norm": 0.9921814925845447,
-      "learning_rate": 1.9137008411438213e-05,
-      "loss": 0.9743,
-      "step": 450
-    },
-    {
-      "epoch": 0.17697548889478806,
-      "grad_norm": 0.9946357977754318,
-      "learning_rate": 1.8889150393715627e-05,
-      "loss": 0.9674,
-      "step": 500
-    },
-    {
-      "epoch": 0.19467303778426687,
-      "grad_norm": 0.936001151695521,
-      "learning_rate": 1.8612085847777215e-05,
-      "loss": 0.9679,
-      "step": 550
-    },
-    {
-      "epoch": 0.2123705866737457,
-      "grad_norm": 0.9739097270294552,
-      "learning_rate": 1.8306725107357933e-05,
-      "loss": 0.9622,
-      "step": 600
-    },
-    {
-      "epoch": 0.2300681355632245,
-      "grad_norm": 1.0799028416482208,
-      "learning_rate": 1.7974071477237887e-05,
-      "loss": 0.976,
-      "step": 650
-    },
-    {
-      "epoch": 0.2477656844527033,
-      "grad_norm": 0.9330950566244366,
-      "learning_rate": 1.7615217936746246e-05,
-      "loss": 0.9567,
-      "step": 700
-    },
-    {
-      "epoch": 0.2654632333421821,
-      "grad_norm": 0.8832817518420761,
-      "learning_rate": 1.7231343548627085e-05,
-      "loss": 0.9523,
-      "step": 750
-    },
-    {
-      "epoch": 0.2831607822316609,
-      "grad_norm": 0.9002124999100791,
-      "learning_rate": 1.6823709585066308e-05,
-      "loss": 0.9466,
-      "step": 800
-    },
-    {
-      "epoch": 0.30085833112113974,
-      "grad_norm": 0.9157647153217237,
-      "learning_rate": 1.6393655383608132e-05,
-      "loss": 0.9429,
-      "step": 850
-    },
-    {
-      "epoch": 0.3185558800106185,
-      "grad_norm": 5.464496113406752,
-      "learning_rate": 1.594259394657707e-05,
-      "loss": 0.9695,
-      "step": 900
-    },
-    {
-      "epoch": 0.33625342890009735,
-      "grad_norm": 0.9883051330334294,
-      "learning_rate": 1.5472007298464117e-05,
-      "loss": 0.944,
-      "step": 950
-    },
-    {
-      "epoch": 0.3539509777895761,
-      "grad_norm": 0.8832131211036971,
-      "learning_rate": 1.4983441616531152e-05,
-      "loss": 0.937,
-      "step": 1000
-    },
-    {
-      "epoch": 0.37164852667905496,
-      "grad_norm": 0.9595043070379917,
-      "learning_rate": 1.4478502150632503e-05,
-      "loss": 0.9313,
-      "step": 1050
-    },
-    {
-      "epoch": 0.38934607556853373,
-      "grad_norm": 1.035782879874017,
-      "learning_rate": 1.3958847948945428e-05,
-      "loss": 0.9264,
-      "step": 1100
-    },
-    {
-      "epoch": 0.40704362445801257,
-      "grad_norm": 0.961077579119761,
-      "learning_rate": 1.3426186406938769e-05,
-      "loss": 0.924,
-      "step": 1150
-    },
-    {
-      "epoch": 0.4247411733474914,
-      "grad_norm": 0.882487785914029,
-      "learning_rate": 1.2882267657489908e-05,
-      "loss": 0.9226,
-      "step": 1200
-    },
-    {
-      "epoch": 0.44243872223697017,
-      "grad_norm": 0.8676230908883836,
-      "learning_rate": 1.2328878820582122e-05,
-      "loss": 0.9134,
-      "step": 1250
-    },
-    {
-      "epoch": 0.460136271126449,
-      "grad_norm": 0.8663581800411811,
-      "learning_rate": 1.1767838131475654e-05,
-      "loss": 0.917,
-      "step": 1300
-    },
-    {
-      "epoch": 0.4778338200159278,
-      "grad_norm": 1.1913367079153512,
-      "learning_rate": 1.1200988966645286e-05,
-      "loss": 0.9181,
-      "step": 1350
-    },
-    {
-      "epoch": 0.4955313689054066,
-      "grad_norm": 0.8563122357771151,
-      "learning_rate": 1.0630193787112994e-05,
-      "loss": 0.903,
-      "step": 1400
-    },
-    {
-      "epoch": 0.5132289177948854,
-      "grad_norm": 0.8518715044091224,
-      "learning_rate": 1.005732801907567e-05,
-      "loss": 0.9085,
-      "step": 1450
-    },
-    {
-      "epoch": 0.5309264666843642,
-      "grad_norm": 0.9224776221871871,
-      "learning_rate": 9.484273891933982e-06,
-      "loss": 0.9086,
-      "step": 1500
-    },
-    {
-      "epoch": 0.548624015573843,
-      "grad_norm": 0.9846231662432429,
-      "learning_rate": 8.912914253968391e-06,
-      "loss": 0.9017,
-      "step": 1550
-    },
-    {
-      "epoch": 0.5663215644633218,
-      "grad_norm": 0.8883898612449065,
-      "learning_rate": 8.345126385981737e-06,
-      "loss": 0.896,
-      "step": 1600
-    },
-    {
-      "epoch": 0.5840191133528007,
-      "grad_norm": 0.9045281001313086,
-      "learning_rate": 7.782775833234522e-06,
-      "loss": 0.9033,
-      "step": 1650
-    },
-    {
-      "epoch": 0.6017166622422795,
-      "grad_norm": 0.8962005631393778,
-      "learning_rate": 7.227710275938987e-06,
-      "loss": 0.8979,
-      "step": 1700
-    },
-    {
-      "epoch": 0.6194142111317582,
-      "grad_norm": 0.8638732399362996,
-      "learning_rate": 6.68175345845119e-06,
-      "loss": 0.8889,
-      "step": 1750
-    },
-    {
-      "epoch": 0.637111760021237,
-      "grad_norm": 1.0900143725646285,
-      "learning_rate": 6.146699197107715e-06,
-      "loss": 0.8957,
-      "step": 1800
-    },
-    {
-      "epoch": 0.6548093089107159,
-      "grad_norm": 0.9103789814206331,
-      "learning_rate": 5.6243054863949675e-06,
-      "loss": 0.8858,
-      "step": 1850
-    },
-    {
-      "epoch": 0.6725068578001947,
-      "grad_norm": 0.9300189349082482,
-      "learning_rate": 5.116288722816087e-06,
-      "loss": 0.8822,
-      "step": 1900
-    },
-    {
-      "epoch": 0.6902044066896735,
-      "grad_norm": 0.8743852772084869,
-      "learning_rate": 4.6243180654337975e-06,
-      "loss": 0.9172,
-      "step": 1950
-    },
-    {
-      "epoch": 0.7079019555791523,
-      "grad_norm": 0.9546014887497036,
-      "learning_rate": 4.1500099516183555e-06,
-      "loss": 1.1497,
-      "step": 2000
-    },
-    {
-      "epoch": 0.7255995044686311,
-      "grad_norm": 0.9614448138630033,
-      "learning_rate": 3.6949227860198712e-06,
-      "loss": 1.0692,
-      "step": 2050
-    },
-    {
-      "epoch": 0.7432970533581099,
-      "grad_norm": 0.848345849582495,
-      "learning_rate": 3.2605518202151577e-06,
-      "loss": 1.0286,
-      "step": 2100
-    },
-    {
-      "epoch": 0.7609946022475887,
-      "grad_norm": 1.0998684999586237,
-      "learning_rate": 2.8483242398526723e-06,
-      "loss": 0.8803,
-      "step": 2150
-    },
-    {
-      "epoch": 0.7786921511370675,
-      "grad_norm": 0.8279471296348094,
-      "learning_rate": 2.4595944754374723e-06,
-      "loss": 0.8885,
-      "step": 2200
-    },
-    {
-      "epoch": 0.7963897000265463,
-      "grad_norm": 0.8602056476985009,
-      "learning_rate": 2.0956397521631666e-06,
-      "loss": 0.8704,
-      "step": 2250
-    },
-    {
-      "epoch": 0.8140872489160251,
-      "grad_norm": 0.9945305020433657,
-      "learning_rate": 1.757655893412622e-06,
-      "loss": 0.8677,
-      "step": 2300
-    },
-    {
-      "epoch": 0.831784797805504,
-      "grad_norm": 0.850430777727199,
-      "learning_rate": 1.4467533917154842e-06,
-      "loss": 0.877,
-      "step": 2350
-    },
-    {
-      "epoch": 0.8494823466949828,
-      "grad_norm": 0.896244091668865,
-      "learning_rate": 1.1639537600719764e-06,
-      "loss": 0.8752,
-      "step": 2400
-    },
-    {
-      "epoch": 0.8671798955844615,
-      "grad_norm": 0.9590853387394762,
-      "learning_rate": 9.101861756312369e-07,
-      "loss": 0.8737,
-      "step": 2450
-    },
-    {
-      "epoch": 0.8848774444739403,
-      "grad_norm": 0.805579311652691,
-      "learning_rate": 6.862844267517643e-07,
-      "loss": 0.8792,
-      "step": 2500
-    },
-    {
-      "epoch": 0.9025749933634192,
-      "grad_norm": 0.8767296521208048,
-      "learning_rate": 4.929841734749063e-07,
-      "loss": 0.875,
-      "step": 2550
-    },
-    {
-      "epoch": 0.920272542252898,
-      "grad_norm": 0.9004753221903812,
-      "learning_rate": 3.309205304124552e-07,
-      "loss": 0.8703,
-      "step": 2600
-    },
-    {
-      "epoch": 0.9379700911423767,
-      "grad_norm": 0.8622876486385493,
-      "learning_rate": 2.0062597999009114e-07,
-      "loss": 0.8815,
-      "step": 2650
-    },
-    {
-      "epoch": 0.9556676400318556,
-      "grad_norm": 0.896417353499626,
-      "learning_rate": 1.0252862290301092e-07,
-      "loss": 0.8715,
-      "step": 2700
-    },
-    {
-      "epoch": 0.9733651889213344,
-      "grad_norm": 0.9142726428404561,
-      "learning_rate": 3.6950771532126004e-08,
-      "loss": 0.8657,
-      "step": 2750
-    },
-    {
-      "epoch": 0.9910627378108132,
-      "grad_norm": 0.8664983124317569,
-      "learning_rate": 4.1078909423253325e-09,
-      "loss": 0.8791,
-      "step": 2800
-    },
-    {
-      "epoch": 0.9999115122555526,
-      "step": 2825,
-      "total_flos": 4.909801179132723e+16,
-      "train_loss": 0.9400995777771536,
-      "train_runtime": 88050.3139,
-      "train_samples_per_second": 0.513,
-      "train_steps_per_second": 0.032
-    }
-  ],
-  "logging_steps": 50,
-  "max_steps": 2825,
-  "num_input_tokens_seen": 0,
-  "num_train_epochs": 1,
-  "save_steps": 500,
-  "total_flos": 4.909801179132723e+16,
-  "train_batch_size": 4,
-  "trial_name": null,
-  "trial_params": null
-}

llavav16bta_cp3d1688_longb_qafull_llhound_fullimg_trainllm_e1_bz256_32f1688/config.json

@@ -1,98 +0,0 @@
-{
-  "architectures": [
-    "LlavaInterleaveForConditionalGeneration"
-  ],
-  "btadapter": true,
-  "btadapter_depth": 4,
-  "clip_post_pretrain": null,
-  "clip_weight": "/group/40043/ruyangliu/weight/CLIP-L334",
-  "frame_shape": [
-    24,
-    24
-  ],
-  "hidden_size": 4096,
-  "ignore_index": -100,
-  "image_grid_pinpoints": [
-    [
-      336,
-      672
-    ],
-    [
-      672,
-      336
-    ],
-    [
-      672,
-      672
-    ],
-    [
-      1008,
-      336
-    ],
-    [
-      336,
-      1008
-    ]
-  ],
-  "image_pooling_kernel": [
-    1,
-    3,
-    3
-  ],
-  "image_pooling_stride": [
-    1,
-    3,
-    3
-  ],
-  "image_token_index": 32000,
-  "long_clip": true,
-  "max_T": 64,
-  "model_type": "llava_next",
-  "pad_token_id": 0,
-  "pllava_pooling_shape": null,
-  "pooling": "clipST_3d",
-  "pooling_kernel": [
-    2,
-    3,
-    3
-  ],
-  "pooling_stride": [
-    2,
-    3,
-    3
-  ],
-  "pooling_temp": 0.01,
-  "projector_hidden_act": "gelu",
-  "text_config": {
-    "_name_or_path": "lmsys/vicuna-7b-v1.5",
-    "architectures": [
-      "LlamaForCausalLM"
-    ],
-    "max_position_embeddings": 4096,
-    "model_type": "llama",
-    "pad_token_id": 0,
-    "rms_norm_eps": 1e-05,
-    "torch_dtype": "float16",
-    "vocab_size": 32064
-  },
-  "torch_dtype": "bfloat16",
-  "transformers_version": "4.40.0",
-  "use_image_newline_parameter": true,
-  "video_input": "mean",
-  "vision_config": {
-    "depth": 4,
-    "hidden_size": 1024,
-    "image_size": 336,
-    "intermediate_size": 4096,
-    "max_T": 64,
-    "model_type": "clip_vision_model",
-    "num_attention_heads": 16,
-    "num_hidden_layers": 24,
-    "patch_size": 14,
-    "projection_dim": 768,
-    "vocab_size": 32000
-  },
-  "vision_feature_layer": -2,
-  "vision_feature_select_strategy": "default",
-  "vocab_size": 32064
-}

llavav16bta_cp3d1688_longb_qafull_llhound_fullimg_trainllm_e1_bz256_32f1688/generation_config.json

@@ -1,7 +0,0 @@
-{
-  "_from_model_config": true,
-  "bos_token_id": 1,
-  "eos_token_id": 2,
-  "pad_token_id": 0,
-  "transformers_version": "4.40.0"
-}

llavav16bta_cp3d1688_longb_qafull_llhound_fullimg_trainllm_e1_bz256_32f1688/model-00001-of-00003.safetensors

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:2468064fc5ed3bc395743532b2d6565ee20a678efb5a292527aa25d80a547574
-size 4980641554

llavav16bta_cp3d1688_longb_qafull_llhound_fullimg_trainllm_e1_bz256_32f1688/trainer_state.json

@@ -1,485 +0,0 @@
-{
-  "best_metric": null,
-  "best_model_checkpoint": null,
-  "epoch": 0.9998461775111521,
-  "eval_steps": 500,
-  "global_step": 3250,
-  "is_hyper_param_search": false,
-  "is_local_process_zero": true,
-  "is_world_process_zero": true,
-  "log_history": [
-    {
-      "epoch": 0.015382248884786957,
-      "grad_norm": 1.2864877128592522,
-      "learning_rate": 1.0204081632653063e-05,
-      "loss": 1.211,
-      "step": 50
-    },
-    {
-      "epoch": 0.030764497769573913,
-      "grad_norm": 1.227813083166002,
-      "learning_rate": 1.999998013186146e-05,
-      "loss": 1.0177,
-      "step": 100
-    },
-    {
-      "epoch": 0.046146746654360866,
-      "grad_norm": 1.16006238153043,
-      "learning_rate": 1.9986572140101312e-05,
-      "loss": 0.9909,
-      "step": 150
-    },
-    {
-      "epoch": 0.06152899553914783,
-      "grad_norm": 1.2120743368126807,
-      "learning_rate": 1.9948367447837503e-05,
-      "loss": 0.9853,
-      "step": 200
-    },
-    {
-      "epoch": 0.07691124442393478,
-      "grad_norm": 1.0434928463049529,
-      "learning_rate": 1.9885460917481137e-05,
-      "loss": 0.9766,
-      "step": 250
-    },
-    {
-      "epoch": 0.09229349330872173,
-      "grad_norm": 1.0799628915416184,
-      "learning_rate": 1.9798008746210894e-05,
-      "loss": 0.9819,
-      "step": 300
-    },
-    {
-      "epoch": 0.10767574219350869,
-      "grad_norm": 1.0390775292248948,
-      "learning_rate": 1.968622807813479e-05,
-      "loss": 0.9678,
-      "step": 350
-    },
-    {
-      "epoch": 0.12305799107829565,
-      "grad_norm": 1.3974223368156584,
-      "learning_rate": 1.955039646512041e-05,
-      "loss": 0.9782,
-      "step": 400
-    },
-    {
-      "epoch": 0.1384402399630826,
-      "grad_norm": 1.1107339192286112,
-      "learning_rate": 1.9390851177632496e-05,
-      "loss": 0.9602,
-      "step": 450
-    },
-    {
-      "epoch": 0.15382248884786956,
-      "grad_norm": 1.0579993349228267,
-      "learning_rate": 1.9207988367288956e-05,
-      "loss": 0.9571,
-      "step": 500
-    },
-    {
-      "epoch": 0.1692047377326565,
-      "grad_norm": 1.1144151233888806,
-      "learning_rate": 1.9002262083214728e-05,
-      "loss": 0.9669,
-      "step": 550
-    },
-    {
-      "epoch": 0.18458698661744347,
-      "grad_norm": 1.187591950721483,
-      "learning_rate": 1.8774183144635894e-05,
-      "loss": 0.958,
-      "step": 600
-    },
-    {
-      "epoch": 0.19996923550223042,
-      "grad_norm": 1.3767826488829,
-      "learning_rate": 1.8524317872513434e-05,
-      "loss": 0.9481,
-      "step": 650
-    },
-    {
-      "epoch": 0.21535148438701737,
-      "grad_norm": 1.0853939205331895,
-      "learning_rate": 1.8253286683365855e-05,
-      "loss": 0.9434,
-      "step": 700
-    },
-    {
-      "epoch": 0.23073373327180433,
-      "grad_norm": 1.0358359930096885,
-      "learning_rate": 1.7961762548772413e-05,
-      "loss": 0.9524,
-      "step": 750
-    },
-    {
-      "epoch": 0.2461159821565913,
-      "grad_norm": 1.0885430781645802,
-      "learning_rate": 1.7650469324381843e-05,
-      "loss": 0.9443,
-      "step": 800
-    },
-    {
-      "epoch": 0.26149823104137826,
-      "grad_norm": 1.139432989706956,
-      "learning_rate": 1.732017995257575e-05,
-      "loss": 0.9398,
-      "step": 850
-    },
-    {
-      "epoch": 0.2768804799261652,
-      "grad_norm": 1.1616765089635157,
-      "learning_rate": 1.697171454324943e-05,
-      "loss": 0.9318,
-      "step": 900
-    },
-    {
-      "epoch": 0.29226272881095217,
-      "grad_norm": 1.102418985915195,
-      "learning_rate": 1.660593833747557e-05,
-      "loss": 0.9348,
-      "step": 950
-    },
-    {
-      "epoch": 0.3076449776957391,
-      "grad_norm": 1.0377834162761999,
-      "learning_rate": 1.6223759559107028e-05,
-      "loss": 0.9206,
-      "step": 1000
-    },
-    {
-      "epoch": 0.3230272265805261,
-      "grad_norm": 1.103549274843009,
-      "learning_rate": 1.5826127159653203e-05,
-      "loss": 0.9353,
-      "step": 1050
-    },
-    {
-      "epoch": 0.338409475465313,
-      "grad_norm": 1.0090406165137933,
-      "learning_rate": 1.5414028462029454e-05,
-      "loss": 0.9363,
-      "step": 1100
-    },
-    {
-      "epoch": 0.3537917243501,
-      "grad_norm": 1.1329344585144494,
-      "learning_rate": 1.4988486709030166e-05,
-      "loss": 0.9347,
-      "step": 1150
-    },
-    {
-      "epoch": 0.36917397323488693,
-      "grad_norm": 1.0307615452725163,
-      "learning_rate": 1.4550558522612584e-05,
-      "loss": 0.9231,
-      "step": 1200
-    },
-    {
-      "epoch": 0.3845562221196739,
-      "grad_norm": 1.1133951229282608,
-      "learning_rate": 1.410133128030009e-05,
-      "loss": 0.9281,
-      "step": 1250
-    },
-    {
-      "epoch": 0.39993847100446084,
-      "grad_norm": 1.0490189419573543,
-      "learning_rate": 1.3641920415219224e-05,
-      "loss": 0.9229,
-      "step": 1300
-    },
-    {
-      "epoch": 0.4153207198892478,
-      "grad_norm": 1.0225887710538208,
-      "learning_rate": 1.3173466646474575e-05,
-      "loss": 0.9115,
-      "step": 1350
-    },
-    {
-      "epoch": 0.43070296877403474,
-      "grad_norm": 1.0818630597333305,
-      "learning_rate": 1.2697133146738452e-05,
-      "loss": 0.9124,
-      "step": 1400
-    },
-    {
-      "epoch": 0.4460852176588217,
-      "grad_norm": 1.1222485167556653,
-      "learning_rate": 1.2214102654088248e-05,
-      "loss": 0.9166,
-      "step": 1450
-    },
-    {
-      "epoch": 0.46146746654360865,
-      "grad_norm": 0.9980409625749838,
-      "learning_rate": 1.1725574535262863e-05,
-      "loss": 0.9047,
-      "step": 1500
-    },
-    {
-      "epoch": 0.4768497154283956,
-      "grad_norm": 1.0397039742249545,
-      "learning_rate": 1.1232761807630032e-05,
-      "loss": 0.9079,
-      "step": 1550
-    },
-    {
-      "epoch": 0.4922319643131826,
-      "grad_norm": 1.0715968646982938,
-      "learning_rate": 1.0736888127259132e-05,
-      "loss": 0.9061,
-      "step": 1600
-    },
-    {
-      "epoch": 0.5076142131979695,
-      "grad_norm": 1.0724763367894545,
-      "learning_rate": 1.023918475057803e-05,
-      "loss": 0.8947,
-      "step": 1650
-    },
-    {
-      "epoch": 0.5229964620827565,
-      "grad_norm": 1.03882946183293,
-      "learning_rate": 9.740887477158194e-06,
-      "loss": 0.9053,
-      "step": 1700
-    },
-    {
-      "epoch": 0.5383787109675434,
-      "grad_norm": 1.0204487030578802,
-      "learning_rate": 9.24323358121919e-06,
-      "loss": 0.9016,
-      "step": 1750
-    },
-    {
-      "epoch": 0.5537609598523304,
-      "grad_norm": 1.0085542891249055,
-      "learning_rate": 8.747458739471603e-06,
-      "loss": 0.8988,
-      "step": 1800
-    },
-    {
-      "epoch": 0.5691432087371173,
-      "grad_norm": 1.1293770623412562,
-      "learning_rate": 8.254793962926579e-06,
-      "loss": 0.8967,
-      "step": 1850
-    },
-    {
-      "epoch": 0.5845254576219043,
-      "grad_norm": 1.1860769177336596,
-      "learning_rate": 7.76646254029032e-06,
-      "loss": 0.8967,
-      "step": 1900
-    },
-    {
-      "epoch": 0.5999077065066912,
-      "grad_norm": 1.062132771926321,
-      "learning_rate": 7.283677000533106e-06,
-      "loss": 0.8936,
-      "step": 1950
-    },
-    {
-      "epoch": 0.6152899553914782,
-      "grad_norm": 1.1679358844534626,
-      "learning_rate": 6.807636102174764e-06,
-      "loss": 0.8928,
-      "step": 2000
-    },
-    {
-      "epoch": 0.6306722042762651,
-      "grad_norm": 1.0610336394985995,
-      "learning_rate": 6.339521856762254e-06,
-      "loss": 0.8845,
-      "step": 2050
-    },
-    {
-      "epoch": 0.6460544531610521,
-      "grad_norm": 1.084745782348564,
-      "learning_rate": 5.880496593930021e-06,
-      "loss": 0.8869,
-      "step": 2100
-    },
-    {
-      "epoch": 0.6614367020458392,
-      "grad_norm": 0.9948936563241075,
-      "learning_rate": 5.431700075330654e-06,
-      "loss": 0.8817,
-      "step": 2150
-    },
-    {
-      "epoch": 0.676818950930626,
-      "grad_norm": 1.0955833872820746,
-      "learning_rate": 4.9942466646019e-06,
-      "loss": 0.8857,
-      "step": 2200
-    },
-    {
-      "epoch": 0.6922011998154131,
-      "grad_norm": 1.0328397346172227,
-      "learning_rate": 4.569222560397126e-06,
-      "loss": 0.8842,
-      "step": 2250
-    },
-    {
-      "epoch": 0.7075834487002,
-      "grad_norm": 1.0439140807601641,
-      "learning_rate": 4.157683099349444e-06,
-      "loss": 0.8797,
-      "step": 2300
-    },
-    {
-      "epoch": 0.722965697584987,
-      "grad_norm": 1.0659188075287298,
-      "learning_rate": 3.760650135666476e-06,
-      "loss": 0.8765,
-      "step": 2350
-    },
-    {
-      "epoch": 0.7383479464697739,
-      "grad_norm": 1.0691950735737183,
-      "learning_rate": 3.3791095038620515e-06,
-      "loss": 0.8715,
-      "step": 2400
-    },
-    {
-      "epoch": 0.7537301953545609,
-      "grad_norm": 1.0882575411106976,
-      "learning_rate": 3.0140085709249666e-06,
-      "loss": 0.8754,
-      "step": 2450
-    },
-    {
-      "epoch": 0.7691124442393478,
-      "grad_norm": 1.1112353348166593,
-      "learning_rate": 2.6662538840027984e-06,
-      "loss": 0.873,
-      "step": 2500
-    },
-    {
-      "epoch": 0.7844946931241348,
-      "grad_norm": 1.1312531360561675,
-      "learning_rate": 2.3367089194415226e-06,
-      "loss": 0.8776,
-      "step": 2550
-    },
-    {
-      "epoch": 0.7998769420089217,
-      "grad_norm": 1.118822941453432,
-      "learning_rate": 2.026191938770128e-06,
-      "loss": 0.8769,
-      "step": 2600
-    },
-    {
-      "epoch": 0.8152591908937087,
-      "grad_norm": 1.0617562102838363,
-      "learning_rate": 1.7354739569538081e-06,
-      "loss": 0.8611,
-      "step": 2650
-    },
-    {
-      "epoch": 0.8306414397784956,
-      "grad_norm": 1.1190891753348724,
-      "learning_rate": 1.4652768279605734e-06,
-      "loss": 0.8778,
-      "step": 2700
-    },
-    {
-      "epoch": 0.8460236886632826,
-      "grad_norm": 1.1025665252752193,
-      "learning_rate": 1.2162714523948104e-06,
-      "loss": 0.8582,
-      "step": 2750
-    },
-    {
-      "epoch": 0.8614059375480695,
-      "grad_norm": 1.0621744666965152,
-      "learning_rate": 9.89076111648284e-07,
-      "loss": 0.8643,
-      "step": 2800
-    },
-    {
-      "epoch": 0.8767881864328565,
-      "grad_norm": 0.9763313630276178,
-      "learning_rate": 7.842549327048366e-07,
-      "loss": 0.8667,
-      "step": 2850
-    },
-    {
-      "epoch": 0.8921704353176434,
-      "grad_norm": 1.0393762329975094,
-      "learning_rate": 6.023164874107202e-07,
-      "loss": 0.8677,
-      "step": 2900
-    },
-    {
-      "epoch": 0.9075526842024304,
-      "grad_norm": 1.1026189621389317,
-      "learning_rate": 4.437125296885847e-07,
-      "loss": 0.8649,
-      "step": 2950
-    },
-    {
-      "epoch": 0.9229349330872173,
-      "grad_norm": 1.073850367410434,
-      "learning_rate": 3.088368738305747e-07,
-      "loss": 0.8771,
-      "step": 3000
-    },
-    {
-      "epoch": 0.9383171819720043,
-      "grad_norm": 1.0327578416174346,
-      "learning_rate": 1.9802441665580208e-07,
-      "loss": 0.8678,
-      "step": 3050
-    },
-    {
-      "epoch": 0.9536994308567912,
-      "grad_norm": 1.035131432577654,
-      "learning_rate": 1.1155030596013638e-07,
-      "loss": 0.8662,
-      "step": 3100
-    },
-    {
-      "epoch": 0.9690816797415782,
-      "grad_norm": 1.0361099744000193,
-      "learning_rate": 4.962925732307278e-08,
-      "loss": 0.8588,
-      "step": 3150
-    },
-    {
-      "epoch": 0.9844639286263652,
-      "grad_norm": 1.087919608094763,
-      "learning_rate": 1.2415020968049585e-08,
-      "loss": 0.8733,
-      "step": 3200
-    },
-    {
-      "epoch": 0.9998461775111521,
-      "grad_norm": 1.0473814351391442,
-      "learning_rate": 0.0,
-      "loss": 0.8751,
-      "step": 3250
-    },
-    {
-      "epoch": 0.9998461775111521,
-      "step": 3250,
-      "total_flos": 1.7053270040117248e+16,
-      "train_loss": 0.9156783118614784,
-      "train_runtime": 46310.2792,
-      "train_samples_per_second": 0.562,
-      "train_steps_per_second": 0.07
-    }
-  ],
-  "logging_steps": 50,
-  "max_steps": 3250,
-  "num_input_tokens_seen": 0,
-  "num_train_epochs": 1,
-  "save_steps": 500,
-  "total_flos": 1.7053270040117248e+16,
-  "train_batch_size": 4,
-  "trial_name": null,
-  "trial_params": null
-}