Model save

README.md

@@ -0,0 +1,61 @@
+---
+library_name: transformers
+license: apache-2.0
+base_model: alignment-handbook/zephyr-7b-sft-full
+tags:
+- trl
+- dpo
+- generated_from_trainer
+model-index:
+- name: zephyr-7b-uf-rc-small-dpo
+  results: []
+---
+
+<!-- This model card has been generated automatically according to the information the Trainer had access to. You
+should probably proofread and complete it, then remove this comment. -->
+
+# zephyr-7b-uf-rc-small-dpo
+
+This model is a fine-tuned version of [alignment-handbook/zephyr-7b-sft-full](https://huggingface.co/alignment-handbook/zephyr-7b-sft-full) on an unknown dataset.
+
+## Model description
+
+More information needed
+
+## Intended uses & limitations
+
+More information needed
+
+## Training and evaluation data
+
+More information needed
+
+## Training procedure
+
+### Training hyperparameters
+
+The following hyperparameters were used during training:
+- learning_rate: 5e-07
+- train_batch_size: 8
+- eval_batch_size: 8
+- seed: 42
+- distributed_type: multi-GPU
+- num_devices: 8
+- gradient_accumulation_steps: 4
+- total_train_batch_size: 256
+- total_eval_batch_size: 64
+- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
+- lr_scheduler_type: cosine
+- lr_scheduler_warmup_ratio: 0.1
+- num_epochs: 1
+
+### Training results
+
+
+
+### Framework versions
+
+- Transformers 4.44.2
+- Pytorch 2.2.0a0+81ea7a4
+- Datasets 2.21.0
+- Tokenizers 0.19.1

all_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 0.9874476987447699,
+    "total_flos": 0.0,
+    "train_loss": 0.47203683449050127,
+    "train_runtime": 1681.7427,
+    "train_samples": 15283,
+    "train_samples_per_second": 9.088,
+    "train_steps_per_second": 0.035
+}

generation_config.json

@@ -0,0 +1,6 @@
+{
+  "_from_model_config": true,
+  "bos_token_id": 1,
+  "eos_token_id": 2,
+  "transformers_version": "4.44.2"
+}

train_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 0.9874476987447699,
+    "total_flos": 0.0,
+    "train_loss": 0.47203683449050127,
+    "train_runtime": 1681.7427,
+    "train_samples": 15283,
+    "train_samples_per_second": 9.088,
+    "train_steps_per_second": 0.035
+}

trainer_state.json

@@ -0,0 +1,132 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 0.9874476987447699,
+  "eval_steps": 100,
+  "global_step": 59,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.016736401673640166,
+      "grad_norm": 7.482011580623352,
+      "learning_rate": 8.333333333333333e-08,
+      "logits/chosen": -2.6849496364593506,
+      "logits/rejected": -2.659409523010254,
+      "logps/chosen": -384.029541015625,
+      "logps/rejected": -397.6531982421875,
+      "loss": 0.6931,
+      "rewards/accuracies": 0.0,
+      "rewards/chosen": 0.0,
+      "rewards/margins": 0.0,
+      "rewards/rejected": 0.0,
+      "step": 1
+    },
+    {
+      "epoch": 0.16736401673640167,
+      "grad_norm": 7.941364250628201,
+      "learning_rate": 4.930057285201027e-07,
+      "logits/chosen": -2.754725456237793,
+      "logits/rejected": -2.730708122253418,
+      "logps/chosen": -380.634521484375,
+      "logps/rejected": -407.98577880859375,
+      "loss": 0.6865,
+      "rewards/accuracies": 0.6527777910232544,
+      "rewards/chosen": 0.006238562986254692,
+      "rewards/margins": 0.013576013036072254,
+      "rewards/rejected": -0.0073374491184949875,
+      "step": 10
+    },
+    {
+      "epoch": 0.33472803347280333,
+      "grad_norm": 9.195790455018836,
+      "learning_rate": 4.187457503795526e-07,
+      "logits/chosen": -2.747068405151367,
+      "logits/rejected": -2.716444492340088,
+      "logps/chosen": -358.1653747558594,
+      "logps/rejected": -392.8216552734375,
+      "loss": 0.617,
+      "rewards/accuracies": 0.862500011920929,
+      "rewards/chosen": 0.05825890973210335,
+      "rewards/margins": 0.16466794908046722,
+      "rewards/rejected": -0.10640902817249298,
+      "step": 20
+    },
+    {
+      "epoch": 0.502092050209205,
+      "grad_norm": 10.217760497962127,
+      "learning_rate": 2.8691164100062034e-07,
+      "logits/chosen": -2.7601308822631836,
+      "logits/rejected": -2.7297308444976807,
+      "logps/chosen": -406.65985107421875,
+      "logps/rejected": -476.3982849121094,
+      "loss": 0.4796,
+      "rewards/accuracies": 0.871874988079071,
+      "rewards/chosen": -0.3961808979511261,
+      "rewards/margins": 0.547135591506958,
+      "rewards/rejected": -0.9433165788650513,
+      "step": 30
+    },
+    {
+      "epoch": 0.6694560669456067,
+      "grad_norm": 12.540869621611261,
+      "learning_rate": 1.4248369943086995e-07,
+      "logits/chosen": -2.7881760597229004,
+      "logits/rejected": -2.7458648681640625,
+      "logps/chosen": -481.26361083984375,
+      "logps/rejected": -644.3431396484375,
+      "loss": 0.3678,
+      "rewards/accuracies": 0.8812500238418579,
+      "rewards/chosen": -1.1410269737243652,
+      "rewards/margins": 1.3094160556793213,
+      "rewards/rejected": -2.4504427909851074,
+      "step": 40
+    },
+    {
+      "epoch": 0.8368200836820083,
+      "grad_norm": 13.132777473342623,
+      "learning_rate": 3.473909705816111e-08,
+      "logits/chosen": -2.7689878940582275,
+      "logits/rejected": -2.7408366203308105,
+      "logps/chosen": -522.9761962890625,
+      "logps/rejected": -689.8461303710938,
+      "loss": 0.3426,
+      "rewards/accuracies": 0.8843749761581421,
+      "rewards/chosen": -1.4519160985946655,
+      "rewards/margins": 1.475476622581482,
+      "rewards/rejected": -2.9273929595947266,
+      "step": 50
+    },
+    {
+      "epoch": 0.9874476987447699,
+      "step": 59,
+      "total_flos": 0.0,
+      "train_loss": 0.47203683449050127,
+      "train_runtime": 1681.7427,
+      "train_samples_per_second": 9.088,
+      "train_steps_per_second": 0.035
+    }
+  ],
+  "logging_steps": 10,
+  "max_steps": 59,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 1,
+  "save_steps": 100,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 0.0,
+  "train_batch_size": 8,
+  "trial_name": null,
+  "trial_params": null
+}