Model save

README.md

@@ -0,0 +1,71 @@
+---
+license: apache-2.0
+base_model: Qwen/Qwen2.5-7B
+tags:
+- generated_from_trainer
+model-index:
+- name: Qwen2.5-7B-spin-10k
+  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. -->
+
+# Qwen2.5-7B-spin-10k
+
+This model is a fine-tuned version of [Qwen/Qwen2.5-7B](https://huggingface.co/Qwen/Qwen2.5-7B) on an unknown dataset.
+It achieves the following results on the evaluation set:
+- Loss: 0.1366
+- Rewards/real: 7.8246
+- Rewards/generated: -2.3247
+- Rewards/accuracies: 1.0
+- Rewards/margins: 10.1492
+- Logps/generated: -263.3269
+- Logps/real: -127.7119
+- Logits/generated: -0.9835
+- Logits/real: -0.6921
+
+## 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: 4
+- eval_batch_size: 4
+- seed: 42
+- distributed_type: multi-GPU
+- num_devices: 4
+- gradient_accumulation_steps: 2
+- total_train_batch_size: 32
+- total_eval_batch_size: 16
+- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
+- lr_scheduler_type: linear
+- lr_scheduler_warmup_ratio: 0.1
+- num_epochs: 1
+
+### Training results
+
+| Training Loss | Epoch  | Step | Validation Loss | Rewards/real | Rewards/generated | Rewards/accuracies | Rewards/margins | Logps/generated | Logps/real | Logits/generated | Logits/real |
+|:-------------:|:------:|:----:|:---------------:|:------------:|:-----------------:|:------------------:|:---------------:|:---------------:|:----------:|:----------------:|:-----------:|
+| 0.1463        | 0.6410 | 50   | 0.1366          | 7.8246       | -2.3247           | 1.0                | 10.1492         | -263.3269       | -127.7119  | -0.9835          | -0.6921     |
+
+
+### Framework versions
+
+- Transformers 4.43.3
+- Pytorch 2.2.2+cu121
+- Datasets 2.20.0
+- Tokenizers 0.19.1

all_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 1.0,
+    "total_flos": 0.0,
+    "train_loss": 0.2174399999471811,
+    "train_runtime": 1268.5918,
+    "train_samples": 2484,
+    "train_samples_per_second": 1.958,
+    "train_steps_per_second": 0.061
+}

generation_config.json

@@ -0,0 +1,6 @@
+{
+  "bos_token_id": 151643,
+  "eos_token_id": 151643,
+  "max_new_tokens": 2048,
+  "transformers_version": "4.43.3"
+}

train_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 1.0,
+    "total_flos": 0.0,
+    "train_loss": 0.2174399999471811,
+    "train_runtime": 1268.5918,
+    "train_samples": 2484,
+    "train_samples_per_second": 1.958,
+    "train_steps_per_second": 0.061
+}

trainer_state.json

@@ -0,0 +1,178 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 1.0,
+  "eval_steps": 50,
+  "global_step": 78,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.01282051282051282,
+      "grad_norm": 169.85406133026603,
+      "learning_rate": 6.25e-08,
+      "logits/generated": -1.1294811964035034,
+      "logits/real": -0.7686214447021484,
+      "logps/generated": -273.1225891113281,
+      "logps/real": -290.2070007324219,
+      "loss": 0.9091,
+      "rewards/accuracies": 0.0,
+      "rewards/generated": 0.0,
+      "rewards/margins": 0.0,
+      "rewards/real": 0.0,
+      "step": 1
+    },
+    {
+      "epoch": 0.1282051282051282,
+      "grad_norm": 63.23414694225829,
+      "learning_rate": 4.857142857142857e-07,
+      "logits/generated": -0.9032017588615417,
+      "logits/real": -0.9559243321418762,
+      "logps/generated": -279.2696838378906,
+      "logps/real": -224.7206268310547,
+      "loss": 0.6755,
+      "rewards/accuracies": 0.8333333134651184,
+      "rewards/generated": -0.39545938372612,
+      "rewards/margins": 0.8560593724250793,
+      "rewards/real": 0.4606000483036041,
+      "step": 10
+    },
+    {
+      "epoch": 0.2564102564102564,
+      "grad_norm": 4.693240657102186,
+      "learning_rate": 4.142857142857143e-07,
+      "logits/generated": -0.951930820941925,
+      "logits/real": -0.8856027722358704,
+      "logps/generated": -284.35430908203125,
+      "logps/real": -156.9384002685547,
+      "loss": 0.1771,
+      "rewards/accuracies": 1.0,
+      "rewards/generated": -1.183829426765442,
+      "rewards/margins": 5.8197126388549805,
+      "rewards/real": 4.635883808135986,
+      "step": 20
+    },
+    {
+      "epoch": 0.38461538461538464,
+      "grad_norm": 1.1912756226499355,
+      "learning_rate": 3.4285714285714286e-07,
+      "logits/generated": -0.5055528283119202,
+      "logits/real": -0.7188040018081665,
+      "logps/generated": -301.6423645019531,
+      "logps/real": -140.4794921875,
+      "loss": 0.1408,
+      "rewards/accuracies": 1.0,
+      "rewards/generated": -2.185486316680908,
+      "rewards/margins": 9.369997024536133,
+      "rewards/real": 7.184512138366699,
+      "step": 30
+    },
+    {
+      "epoch": 0.5128205128205128,
+      "grad_norm": 1.2464411282623236,
+      "learning_rate": 2.714285714285714e-07,
+      "logits/generated": -0.5057858228683472,
+      "logits/real": -0.5780868530273438,
+      "logps/generated": -299.8144226074219,
+      "logps/real": -154.14199829101562,
+      "loss": 0.1457,
+      "rewards/accuracies": 1.0,
+      "rewards/generated": -2.8365063667297363,
+      "rewards/margins": 10.34560489654541,
+      "rewards/real": 7.509098052978516,
+      "step": 40
+    },
+    {
+      "epoch": 0.6410256410256411,
+      "grad_norm": 1.0548152399288286,
+      "learning_rate": 2e-07,
+      "logits/generated": -0.38232049345970154,
+      "logits/real": -0.6496118307113647,
+      "logps/generated": -303.3151550292969,
+      "logps/real": -143.59616088867188,
+      "loss": 0.1463,
+      "rewards/accuracies": 1.0,
+      "rewards/generated": -2.8232903480529785,
+      "rewards/margins": 10.742189407348633,
+      "rewards/real": 7.9188995361328125,
+      "step": 50
+    },
+    {
+      "epoch": 0.6410256410256411,
+      "eval_logits/generated": -0.9835317730903625,
+      "eval_logits/real": -0.6921402812004089,
+      "eval_logps/generated": -263.3268737792969,
+      "eval_logps/real": -127.7119369506836,
+      "eval_loss": 0.13658176362514496,
+      "eval_rewards/accuracies": 1.0,
+      "eval_rewards/generated": -2.3246724605560303,
+      "eval_rewards/margins": 10.149224281311035,
+      "eval_rewards/real": 7.824552059173584,
+      "eval_runtime": 10.8908,
+      "eval_samples_per_second": 4.591,
+      "eval_steps_per_second": 0.367,
+      "step": 50
+    },
+    {
+      "epoch": 0.7692307692307693,
+      "grad_norm": 1.1141881809347227,
+      "learning_rate": 1.2857142857142855e-07,
+      "logits/generated": -0.503508448600769,
+      "logits/real": -0.6755964159965515,
+      "logps/generated": -294.5934143066406,
+      "logps/real": -133.47311401367188,
+      "loss": 0.1358,
+      "rewards/accuracies": 1.0,
+      "rewards/generated": -3.0683863162994385,
+      "rewards/margins": 10.883561134338379,
+      "rewards/real": 7.815173149108887,
+      "step": 60
+    },
+    {
+      "epoch": 0.8974358974358975,
+      "grad_norm": 1.0596728323269835,
+      "learning_rate": 5.714285714285714e-08,
+      "logits/generated": -0.6190778017044067,
+      "logits/real": -0.6327217221260071,
+      "logps/generated": -289.4383850097656,
+      "logps/real": -148.02481079101562,
+      "loss": 0.145,
+      "rewards/accuracies": 1.0,
+      "rewards/generated": -2.6634089946746826,
+      "rewards/margins": 10.635811805725098,
+      "rewards/real": 7.972402095794678,
+      "step": 70
+    },
+    {
+      "epoch": 1.0,
+      "step": 78,
+      "total_flos": 0.0,
+      "train_loss": 0.2174399999471811,
+      "train_runtime": 1268.5918,
+      "train_samples_per_second": 1.958,
+      "train_steps_per_second": 0.061
+    }
+  ],
+  "logging_steps": 10,
+  "max_steps": 78,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 1,
+  "save_steps": 50,
+  "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": 4,
+  "trial_name": null,
+  "trial_params": null
+}