Model save

README.md

@@ -18,6 +18,8 @@ should probably proofread and complete it, then remove this comment. -->
 # SFT-Qwen2.5-Coder-3B_long_v1
 
 This model is a fine-tuned version of [Qwen/Qwen2.5-Coder-3B-Instruct](https://huggingface.co/Qwen/Qwen2.5-Coder-3B-Instruct) on the None dataset.
+It achieves the following results on the evaluation set:
+- Loss: 0.7569
 
 ## Model description
 
@@ -38,7 +40,7 @@ More information needed
 The following hyperparameters were used during training:
 - learning_rate: 0.0001
 - train_batch_size: 1
-- eval_batch_size: 8
+- eval_batch_size: 1
 - seed: 42
 - gradient_accumulation_steps: 8
 - total_train_batch_size: 8
@@ -47,6 +49,22 @@ The following hyperparameters were used during training:
 - lr_scheduler_warmup_ratio: 0.03
 - num_epochs: 3
 
+### Training results
+
+| Training Loss | Epoch  | Step | Validation Loss |
+|:-------------:|:------:|:----:|:---------------:|
+| 0.9906        | 0.2807 | 20   | 0.9487          |
+| 0.8528        | 0.5614 | 40   | 0.8620          |
+| 0.8721        | 0.8421 | 60   | 0.8238          |
+| 0.8059        | 1.1123 | 80   | 0.8018          |
+| 0.8141        | 1.3930 | 100  | 0.7868          |
+| 0.7353        | 1.6737 | 120  | 0.7767          |
+| 0.6779        | 1.9544 | 140  | 0.7647          |
+| 0.6273        | 2.2246 | 160  | 0.7629          |
+| 0.6983        | 2.5053 | 180  | 0.7597          |
+| 0.6958        | 2.7860 | 200  | 0.7569          |
+
+
 ### Framework versions
 
 - PEFT 0.18.0

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:64150ddf1b93e1201ed89104b4610933cb5157950a32dbc4064c780f61df1545
+oid sha256:1b1bb12b446b82d4cdb5000895076f8c66e6544e2d36eaf932edcacf8b081d64
 size 239536272

trainer_state.json

@@ -1,20 +1,180 @@
 {
-  "best_global_step": null,
-  "best_metric": null,
-  "best_model_checkpoint": null,
-  "epoch": 0.2807017543859649,
+  "best_global_step": 200,
+  "best_metric": 0.7568597793579102,
+  "best_model_checkpoint": "j05hr3d/SFT-Qwen2.5-Coder-3B_long_v1/checkpoint-200",
+  "epoch": 3.0,
   "eval_steps": 20,
-  "global_step": 20,
+  "global_step": 216,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
       "epoch": 0.2807017543859649,
-      "grad_norm": 0.47223395109176636,
+      "grad_norm": 0.47658494114875793,
       "learning_rate": 9.425837320574164e-05,
-      "loss": 0.9903,
+      "loss": 0.9906,
       "step": 20
+    },
+    {
+      "epoch": 0.2807017543859649,
+      "eval_loss": 0.9487136006355286,
+      "eval_runtime": 14.7059,
+      "eval_samples_per_second": 4.352,
+      "eval_steps_per_second": 4.352,
+      "step": 20
+    },
+    {
+      "epoch": 0.5614035087719298,
+      "grad_norm": 0.4057779908180237,
+      "learning_rate": 8.468899521531101e-05,
+      "loss": 0.8528,
+      "step": 40
+    },
+    {
+      "epoch": 0.5614035087719298,
+      "eval_loss": 0.8620332479476929,
+      "eval_runtime": 14.5037,
+      "eval_samples_per_second": 4.413,
+      "eval_steps_per_second": 4.413,
+      "step": 40
+    },
+    {
+      "epoch": 0.8421052631578947,
+      "grad_norm": 0.5583078861236572,
+      "learning_rate": 7.511961722488039e-05,
+      "loss": 0.8721,
+      "step": 60
+    },
+    {
+      "epoch": 0.8421052631578947,
+      "eval_loss": 0.8238226175308228,
+      "eval_runtime": 14.7369,
+      "eval_samples_per_second": 4.343,
+      "eval_steps_per_second": 4.343,
+      "step": 60
+    },
+    {
+      "epoch": 1.1122807017543859,
+      "grad_norm": 0.4012566804885864,
+      "learning_rate": 6.555023923444976e-05,
+      "loss": 0.8059,
+      "step": 80
+    },
+    {
+      "epoch": 1.1122807017543859,
+      "eval_loss": 0.8017938137054443,
+      "eval_runtime": 14.6894,
+      "eval_samples_per_second": 4.357,
+      "eval_steps_per_second": 4.357,
+      "step": 80
+    },
+    {
+      "epoch": 1.3929824561403508,
+      "grad_norm": 0.7860776782035828,
+      "learning_rate": 5.5980861244019145e-05,
+      "loss": 0.8141,
+      "step": 100
+    },
+    {
+      "epoch": 1.3929824561403508,
+      "eval_loss": 0.7868363857269287,
+      "eval_runtime": 14.7,
+      "eval_samples_per_second": 4.354,
+      "eval_steps_per_second": 4.354,
+      "step": 100
+    },
+    {
+      "epoch": 1.6736842105263157,
+      "grad_norm": 0.48762819170951843,
+      "learning_rate": 4.641148325358852e-05,
+      "loss": 0.7353,
+      "step": 120
+    },
+    {
+      "epoch": 1.6736842105263157,
+      "eval_loss": 0.7766555547714233,
+      "eval_runtime": 14.7914,
+      "eval_samples_per_second": 4.327,
+      "eval_steps_per_second": 4.327,
+      "step": 120
+    },
+    {
+      "epoch": 1.9543859649122806,
+      "grad_norm": 0.5413577556610107,
+      "learning_rate": 3.6842105263157895e-05,
+      "loss": 0.6779,
+      "step": 140
+    },
+    {
+      "epoch": 1.9543859649122806,
+      "eval_loss": 0.7647180557250977,
+      "eval_runtime": 14.7626,
+      "eval_samples_per_second": 4.335,
+      "eval_steps_per_second": 4.335,
+      "step": 140
+    },
+    {
+      "epoch": 2.2245614035087717,
+      "grad_norm": 0.8345220685005188,
+      "learning_rate": 2.7272727272727273e-05,
+      "loss": 0.6273,
+      "step": 160
+    },
+    {
+      "epoch": 2.2245614035087717,
+      "eval_loss": 0.7629338502883911,
+      "eval_runtime": 14.7511,
+      "eval_samples_per_second": 4.339,
+      "eval_steps_per_second": 4.339,
+      "step": 160
+    },
+    {
+      "epoch": 2.5052631578947366,
+      "grad_norm": 0.5914349555969238,
+      "learning_rate": 1.770334928229665e-05,
+      "loss": 0.6983,
+      "step": 180
+    },
+    {
+      "epoch": 2.5052631578947366,
+      "eval_loss": 0.7597119808197021,
+      "eval_runtime": 14.7516,
+      "eval_samples_per_second": 4.339,
+      "eval_steps_per_second": 4.339,
+      "step": 180
+    },
+    {
+      "epoch": 2.7859649122807015,
+      "grad_norm": 0.5911722183227539,
+      "learning_rate": 8.133971291866028e-06,
+      "loss": 0.6958,
+      "step": 200
+    },
+    {
+      "epoch": 2.7859649122807015,
+      "eval_loss": 0.7568597793579102,
+      "eval_runtime": 14.6958,
+      "eval_samples_per_second": 4.355,
+      "eval_steps_per_second": 4.355,
+      "step": 200
+    },
+    {
+      "epoch": 3.0,
+      "step": 216,
+      "total_flos": 2.684496203948851e+16,
+      "train_loss": 0.7727184339805886,
+      "train_runtime": 1493.2791,
+      "train_samples_per_second": 1.145,
+      "train_steps_per_second": 0.145
+    },
+    {
+      "epoch": 3.0,
+      "eval_loss": 0.7568597793579102,
+      "eval_runtime": 14.794,
+      "eval_samples_per_second": 4.326,
+      "eval_steps_per_second": 4.326,
+      "step": 216
     }
   ],
   "logging_steps": 20,
@@ -37,13 +197,13 @@
         "should_epoch_stop": false,
         "should_evaluate": false,
         "should_log": false,
-        "should_save": false,
-        "should_training_stop": false
+        "should_save": true,
+        "should_training_stop": true
       },
       "attributes": {}
     }
   },
-  "total_flos": 2320839556472832.0,
+  "total_flos": 2.684496203948851e+16,
   "train_batch_size": 1,
   "trial_name": null,
   "trial_params": null