Model save

README.md

@@ -0,0 +1,58 @@
+---
+base_model: meta-llama/Meta-Llama-3.1-8B-Instruct
+library_name: transformers
+model_name: llama_openthoughts_sorted_sft
+tags:
+- generated_from_trainer
+- trl
+- sft
+licence: license
+---
+
+# Model Card for llama_openthoughts_sorted_sft
+
+This model is a fine-tuned version of [meta-llama/Meta-Llama-3.1-8B-Instruct](https://huggingface.co/meta-llama/Meta-Llama-3.1-8B-Instruct).
+It has been trained using [TRL](https://github.com/huggingface/trl).
+
+## Quick start
+
+```python
+from transformers import pipeline
+
+question = "If you had a time machine, but could only go to the past or the future once and never return, which would you choose and why?"
+generator = pipeline("text-generation", model="imdatta0/llama_openthoughts_sorted_sft", device="cuda")
+output = generator([{"role": "user", "content": question}], max_new_tokens=128, return_full_text=False)[0]
+print(output["generated_text"])
+```
+
+## Training procedure
+
+[<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="150" height="24"/>](https://wandb.ai/imdatta0/reasoning/runs/pk2rom3s) 
+
+
+This model was trained with SFT.
+
+### Framework versions
+
+- TRL: 0.16.0
+- Transformers: 4.50.3
+- Pytorch: 2.5.1
+- Datasets: 3.2.0
+- Tokenizers: 0.21.0
+
+## Citations
+
+
+
+Cite TRL as:
+    
+```bibtex
+@misc{vonwerra2022trl,
+	title        = {{TRL: Transformer Reinforcement Learning}},
+	author       = {Leandro von Werra and Younes Belkada and Lewis Tunstall and Edward Beeching and Tristan Thrush and Nathan Lambert and Shengyi Huang and Kashif Rasul and Quentin Gallouédec},
+	year         = 2020,
+	journal      = {GitHub repository},
+	publisher    = {GitHub},
+	howpublished = {\url{https://github.com/huggingface/trl}}
+}
+```

all_results.json

@@ -0,0 +1,9 @@
+{
+    "total_flos": 8.660994241796768e+18,
+    "train_loss": 0.5592288087014439,
+    "train_runtime": 64181.8949,
+    "train_samples": 113829,
+    "train_samples_per_second": 0.586,
+    "train_steps_per_second": 0.005,
+    "val_samples": 128
+}

config.json

@@ -0,0 +1,39 @@
+{
+  "architectures": [
+    "LlamaForCausalLM"
+  ],
+  "attention_bias": false,
+  "attention_dropout": 0.0,
+  "bos_token_id": 128000,
+  "eos_token_id": [
+    128001,
+    128008,
+    128009
+  ],
+  "head_dim": 128,
+  "hidden_act": "silu",
+  "hidden_size": 4096,
+  "initializer_range": 0.02,
+  "intermediate_size": 14336,
+  "max_position_embeddings": 131072,
+  "mlp_bias": false,
+  "model_type": "llama",
+  "num_attention_heads": 32,
+  "num_hidden_layers": 32,
+  "num_key_value_heads": 8,
+  "pretraining_tp": 1,
+  "rms_norm_eps": 1e-05,
+  "rope_scaling": {
+    "factor": 8.0,
+    "high_freq_factor": 4.0,
+    "low_freq_factor": 1.0,
+    "original_max_position_embeddings": 8192,
+    "rope_type": "llama3"
+  },
+  "rope_theta": 500000.0,
+  "tie_word_embeddings": false,
+  "torch_dtype": "float32",
+  "transformers_version": "4.50.3",
+  "use_cache": true,
+  "vocab_size": 128256
+}

generation_config.json

@@ -0,0 +1,12 @@
+{
+  "bos_token_id": 128000,
+  "do_sample": true,
+  "eos_token_id": [
+    128001,
+    128008,
+    128009
+  ],
+  "temperature": 0.6,
+  "top_p": 0.9,
+  "transformers_version": "4.50.3"
+}

train_results.json

@@ -0,0 +1,9 @@
+{
+    "total_flos": 8.660994241796768e+18,
+    "train_loss": 0.5592288087014439,
+    "train_runtime": 64181.8949,
+    "train_samples": 113829,
+    "train_samples_per_second": 0.586,
+    "train_steps_per_second": 0.005,
+    "val_samples": 128
+}

trainer_state.json

@@ -0,0 +1,260 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 1.0,
+  "eval_steps": 32,
+  "global_step": 294,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.05442176870748299,
+      "grad_norm": 1.1490706205368042,
+      "learning_rate": 2.6666666666666667e-05,
+      "loss": 0.8045,
+      "num_tokens": 41880570.0,
+      "step": 16
+    },
+    {
+      "epoch": 0.10884353741496598,
+      "grad_norm": 1.5008639097213745,
+      "learning_rate": 4.999291986732823e-05,
+      "loss": 0.665,
+      "step": 32
+    },
+    {
+      "epoch": 0.10884353741496598,
+      "eval_loss": 0.651832640171051,
+      "eval_num_tokens": 83689729.0,
+      "eval_runtime": 23.5652,
+      "eval_samples_per_second": 2.122,
+      "eval_steps_per_second": 0.552,
+      "step": 32
+    },
+    {
+      "epoch": 0.16326530612244897,
+      "grad_norm": 1.030198335647583,
+      "learning_rate": 4.942867164927899e-05,
+      "loss": 0.6176,
+      "num_tokens": 125570174.0,
+      "step": 48
+    },
+    {
+      "epoch": 0.21768707482993196,
+      "grad_norm": 1.0640579462051392,
+      "learning_rate": 4.798150758954164e-05,
+      "loss": 0.5868,
+      "step": 64
+    },
+    {
+      "epoch": 0.21768707482993196,
+      "eval_loss": 0.5904918909072876,
+      "eval_num_tokens": 167442808.0,
+      "eval_runtime": 18.7144,
+      "eval_samples_per_second": 2.672,
+      "eval_steps_per_second": 0.695,
+      "step": 64
+    },
+    {
+      "epoch": 0.272108843537415,
+      "grad_norm": 1.4152488708496094,
+      "learning_rate": 4.570373196778427e-05,
+      "loss": 0.5753,
+      "num_tokens": 209308289.0,
+      "step": 80
+    },
+    {
+      "epoch": 0.32653061224489793,
+      "grad_norm": 0.6098703145980835,
+      "learning_rate": 4.267766952966369e-05,
+      "loss": 0.561,
+      "step": 96
+    },
+    {
+      "epoch": 0.32653061224489793,
+      "eval_loss": 0.5666875839233398,
+      "eval_num_tokens": 251165748.0,
+      "eval_runtime": 40.4906,
+      "eval_samples_per_second": 1.235,
+      "eval_steps_per_second": 0.321,
+      "step": 96
+    },
+    {
+      "epoch": 0.38095238095238093,
+      "grad_norm": 0.9113028049468994,
+      "learning_rate": 3.901269005602235e-05,
+      "loss": 0.5498,
+      "num_tokens": 293026285.0,
+      "step": 112
+    },
+    {
+      "epoch": 0.43537414965986393,
+      "grad_norm": 0.38898423314094543,
+      "learning_rate": 3.484125545404854e-05,
+      "loss": 0.5403,
+      "step": 128
+    },
+    {
+      "epoch": 0.43537414965986393,
+      "eval_loss": 0.552385151386261,
+      "eval_num_tokens": 334905303.0,
+      "eval_runtime": 18.667,
+      "eval_samples_per_second": 2.679,
+      "eval_steps_per_second": 0.696,
+      "step": 128
+    },
+    {
+      "epoch": 0.4897959183673469,
+      "grad_norm": 0.41057485342025757,
+      "learning_rate": 3.0314132238824415e-05,
+      "loss": 0.5421,
+      "num_tokens": 376761979.0,
+      "step": 144
+    },
+    {
+      "epoch": 0.54421768707483,
+      "grad_norm": 0.3121841251850128,
+      "learning_rate": 2.5594942438652688e-05,
+      "loss": 0.5327,
+      "step": 160
+    },
+    {
+      "epoch": 0.54421768707483,
+      "eval_loss": 0.5410001873970032,
+      "eval_num_tokens": 418605392.0,
+      "eval_runtime": 18.6226,
+      "eval_samples_per_second": 2.685,
+      "eval_steps_per_second": 0.698,
+      "step": 160
+    },
+    {
+      "epoch": 0.5986394557823129,
+      "grad_norm": 0.3161534070968628,
+      "learning_rate": 2.085424986864065e-05,
+      "loss": 0.5245,
+      "num_tokens": 460446289.0,
+      "step": 176
+    },
+    {
+      "epoch": 0.6530612244897959,
+      "grad_norm": 0.3147275149822235,
+      "learning_rate": 1.6263395510022543e-05,
+      "loss": 0.52,
+      "step": 192
+    },
+    {
+      "epoch": 0.6530612244897959,
+      "eval_loss": 0.5329579710960388,
+      "eval_num_tokens": 502379615.0,
+      "eval_runtime": 18.5679,
+      "eval_samples_per_second": 2.693,
+      "eval_steps_per_second": 0.7,
+      "step": 192
+    },
+    {
+      "epoch": 0.7074829931972789,
+      "grad_norm": 0.24113914370536804,
+      "learning_rate": 1.1988304800673197e-05,
+      "loss": 0.514,
+      "num_tokens": 544263087.0,
+      "step": 208
+    },
+    {
+      "epoch": 0.7619047619047619,
+      "grad_norm": 0.2218102216720581,
+      "learning_rate": 8.183490657468688e-06,
+      "loss": 0.5106,
+      "step": 224
+    },
+    {
+      "epoch": 0.7619047619047619,
+      "eval_loss": 0.5261243581771851,
+      "eval_num_tokens": 586183588.0,
+      "eval_runtime": 18.6928,
+      "eval_samples_per_second": 2.675,
+      "eval_steps_per_second": 0.695,
+      "step": 224
+    },
+    {
+      "epoch": 0.8163265306122449,
+      "grad_norm": 0.2193211168050766,
+      "learning_rate": 4.986468976890993e-06,
+      "loss": 0.5177,
+      "num_tokens": 627943986.0,
+      "step": 240
+    },
+    {
+      "epoch": 0.8707482993197279,
+      "grad_norm": 0.19595304131507874,
+      "learning_rate": 2.512788452234921e-06,
+      "loss": 0.5101,
+      "step": 256
+    },
+    {
+      "epoch": 0.8707482993197279,
+      "eval_loss": 0.5228561162948608,
+      "eval_num_tokens": 669801067.0,
+      "eval_runtime": 18.6498,
+      "eval_samples_per_second": 2.681,
+      "eval_steps_per_second": 0.697,
+      "step": 256
+    },
+    {
+      "epoch": 0.9251700680272109,
+      "grad_norm": 0.17962999641895294,
+      "learning_rate": 8.51854342773295e-07,
+      "loss": 0.5061,
+      "num_tokens": 711714291.0,
+      "step": 272
+    },
+    {
+      "epoch": 0.9795918367346939,
+      "grad_norm": 0.17367339134216309,
+      "learning_rate": 6.369713474366212e-08,
+      "loss": 0.5074,
+      "step": 288
+    },
+    {
+      "epoch": 0.9795918367346939,
+      "eval_loss": 0.5216580629348755,
+      "eval_num_tokens": 753615032.0,
+      "eval_runtime": 18.6674,
+      "eval_samples_per_second": 2.678,
+      "eval_steps_per_second": 0.696,
+      "step": 288
+    },
+    {
+      "epoch": 1.0,
+      "num_tokens": 768687980.0,
+      "step": 294,
+      "total_flos": 8.660994241796768e+18,
+      "train_loss": 0.5592288087014439,
+      "train_runtime": 64181.8949,
+      "train_samples_per_second": 0.586,
+      "train_steps_per_second": 0.005
+    }
+  ],
+  "logging_steps": 16,
+  "max_steps": 294,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 1,
+  "save_steps": 32,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 8.660994241796768e+18,
+  "train_batch_size": 1,
+  "trial_name": null,
+  "trial_params": null
+}