Model save

README.md

@@ -0,0 +1,65 @@
+---
+library_name: transformers
+tags:
+- trl
+- sft
+- generated_from_trainer
+datasets:
+- generator
+model-index:
+- name: llama-3.1-8b-sft-full_noisy_0.5_bon-sft
+  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. -->
+
+# llama-3.1-8b-sft-full_noisy_0.5_bon-sft
+
+This model was trained from scratch on the generator dataset.
+It achieves the following results on the evaluation set:
+- Loss: 1.3351
+
+## 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: 2e-05
+- train_batch_size: 4
+- eval_batch_size: 4
+- seed: 42
+- distributed_type: multi-GPU
+- num_devices: 8
+- total_train_batch_size: 32
+- total_eval_batch_size: 32
+- optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
+- lr_scheduler_type: cosine
+- lr_scheduler_warmup_ratio: 0.1
+- num_epochs: 1
+
+### Training results
+
+| Training Loss | Epoch | Step | Validation Loss |
+|:-------------:|:-----:|:----:|:---------------:|
+| 1.3482        | 1.0   | 58   | 1.3351          |
+
+
+### Framework versions
+
+- Transformers 4.49.0
+- Pytorch 2.6.0+cu124
+- Datasets 3.3.2
+- Tokenizers 0.21.0

all_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 1.0,
+    "total_flos": 12144020029440.0,
+    "train_loss": 1.3671774535343564,
+    "train_runtime": 417.662,
+    "train_samples": 18000,
+    "train_samples_per_second": 4.437,
+    "train_steps_per_second": 0.139
+}

generation_config.json

@@ -0,0 +1,9 @@
+{
+  "_from_model_config": true,
+  "bos_token_id": 128000,
+  "do_sample": true,
+  "eos_token_id": 128001,
+  "temperature": 0.6,
+  "top_p": 0.9,
+  "transformers_version": "4.49.0"
+}

train_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 1.0,
+    "total_flos": 12144020029440.0,
+    "train_loss": 1.3671774535343564,
+    "train_runtime": 417.662,
+    "train_samples": 18000,
+    "train_samples_per_second": 4.437,
+    "train_steps_per_second": 0.139
+}

trainer_state.json

@@ -0,0 +1,134 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 1.0,
+  "eval_steps": 500,
+  "global_step": 58,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.017241379310344827,
+      "grad_norm": 3.1579799955399586,
+      "learning_rate": 3.3333333333333333e-06,
+      "loss": 1.51,
+      "step": 1
+    },
+    {
+      "epoch": 0.08620689655172414,
+      "grad_norm": 3.379171357161518,
+      "learning_rate": 1.6666666666666667e-05,
+      "loss": 1.5004,
+      "step": 5
+    },
+    {
+      "epoch": 0.1724137931034483,
+      "grad_norm": 1.4059604691430452,
+      "learning_rate": 1.9709418174260523e-05,
+      "loss": 1.436,
+      "step": 10
+    },
+    {
+      "epoch": 0.25862068965517243,
+      "grad_norm": 1.5022419061252104,
+      "learning_rate": 1.8557812723014476e-05,
+      "loss": 1.3965,
+      "step": 15
+    },
+    {
+      "epoch": 0.3448275862068966,
+      "grad_norm": 0.992135201683974,
+      "learning_rate": 1.6631226582407954e-05,
+      "loss": 1.3732,
+      "step": 20
+    },
+    {
+      "epoch": 0.43103448275862066,
+      "grad_norm": 0.9997405084274987,
+      "learning_rate": 1.410412805452757e-05,
+      "loss": 1.3528,
+      "step": 25
+    },
+    {
+      "epoch": 0.5172413793103449,
+      "grad_norm": 0.9176842130196954,
+      "learning_rate": 1.1205366802553231e-05,
+      "loss": 1.351,
+      "step": 30
+    },
+    {
+      "epoch": 0.603448275862069,
+      "grad_norm": 0.8809731900228461,
+      "learning_rate": 8.197449621860944e-06,
+      "loss": 1.3364,
+      "step": 35
+    },
+    {
+      "epoch": 0.6896551724137931,
+      "grad_norm": 0.8970095215730851,
+      "learning_rate": 5.352768279562315e-06,
+      "loss": 1.3256,
+      "step": 40
+    },
+    {
+      "epoch": 0.7758620689655172,
+      "grad_norm": 1.0070554966493785,
+      "learning_rate": 2.9289321881345257e-06,
+      "loss": 1.3695,
+      "step": 45
+    },
+    {
+      "epoch": 0.8620689655172413,
+      "grad_norm": 0.8193516116184061,
+      "learning_rate": 1.1454397434679022e-06,
+      "loss": 1.3321,
+      "step": 50
+    },
+    {
+      "epoch": 0.9482758620689655,
+      "grad_norm": 0.8167441937574224,
+      "learning_rate": 1.6380093052856482e-07,
+      "loss": 1.3482,
+      "step": 55
+    },
+    {
+      "epoch": 1.0,
+      "eval_loss": 1.3351141214370728,
+      "eval_runtime": 10.4632,
+      "eval_samples_per_second": 19.115,
+      "eval_steps_per_second": 0.669,
+      "step": 58
+    },
+    {
+      "epoch": 1.0,
+      "step": 58,
+      "total_flos": 12144020029440.0,
+      "train_loss": 1.3671774535343564,
+      "train_runtime": 417.662,
+      "train_samples_per_second": 4.437,
+      "train_steps_per_second": 0.139
+    }
+  ],
+  "logging_steps": 5,
+  "max_steps": 58,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 1,
+  "save_steps": 1500,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 12144020029440.0,
+  "train_batch_size": 4,
+  "trial_name": null,
+  "trial_params": null
+}