End of training

README.md

@@ -4,6 +4,7 @@ license: apache-2.0
 base_model: Qwen/Qwen2.5-32B
 tags:
 - llama-factory
+- full
 - generated_from_trainer
 model-index:
 - name: pretrain
@@ -15,9 +16,9 @@ should probably proofread and complete it, then remove this comment. -->
 
 # pretrain
 
-This model is a fine-tuned version of [Qwen/Qwen2.5-32B](https://huggingface.co/Qwen/Qwen2.5-32B) on an unknown dataset.
+This model is a fine-tuned version of [Qwen/Qwen2.5-32B](https://huggingface.co/Qwen/Qwen2.5-32B) on the openalex_small dataset.
 It achieves the following results on the evaluation set:
-- Loss: 1.1995
+- Loss: 1.1904
 
 ## Model description
 

all_results.json

@@ -1,13 +1,13 @@
 {
-    "epoch": 0.7619047619047619,
-    "eval_loss": 1.4561185836791992,
-    "eval_runtime": 2.8429,
-    "eval_samples_per_second": 5.98,
-    "eval_steps_per_second": 1.055,
-    "perplexity": 4.289278700316188,
-    "total_flos": 2446118092800.0,
-    "train_loss": 1.5153563022613525,
-    "train_runtime": 280.4756,
-    "train_samples_per_second": 0.578,
-    "train_steps_per_second": 0.007
+    "epoch": 0.9454545454545454,
+    "eval_loss": 1.1903643608093262,
+    "eval_runtime": 40.1525,
+    "eval_samples_per_second": 2.441,
+    "eval_steps_per_second": 0.324,
+    "perplexity": 3.2882791091731414,
+    "total_flos": 45330292801536.0,
+    "train_loss": 1.3396454407618597,
+    "train_runtime": 2207.3438,
+    "train_samples_per_second": 0.398,
+    "train_steps_per_second": 0.006
 }

eval_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 2.8,
-    "eval_loss": 1.4561185836791992,
-    "eval_runtime": 2.8429,
-    "eval_samples_per_second": 5.98,
-    "eval_steps_per_second": 1.055,
-    "perplexity": 4.289278700316188
+    "epoch": 0.9454545454545454,
+    "eval_loss": 1.1903643608093262,
+    "eval_runtime": 40.1525,
+    "eval_samples_per_second": 2.441,
+    "eval_steps_per_second": 0.324,
+    "perplexity": 3.2882791091731414
 }

train_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 0.7619047619047619,
-    "total_flos": 2446118092800.0,
-    "train_loss": 1.5153563022613525,
-    "train_runtime": 280.4756,
-    "train_samples_per_second": 0.578,
-    "train_steps_per_second": 0.007
+    "epoch": 0.9454545454545454,
+    "total_flos": 45330292801536.0,
+    "train_loss": 1.3396454407618597,
+    "train_runtime": 2207.3438,
+    "train_samples_per_second": 0.398,
+    "train_steps_per_second": 0.006
 }

trainer_state.json

@@ -1,42 +1,127 @@
 {
   "best_metric": null,
   "best_model_checkpoint": null,
-  "epoch": 0.7619047619047619,
-  "eval_steps": 500,
-  "global_step": 2,
+  "epoch": 0.9454545454545454,
+  "eval_steps": 10,
+  "global_step": 13,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 0.38095238095238093,
-      "grad_norm": 0.5249602161808266,
-      "learning_rate": 0.0001,
-      "loss": 1.5248,
+      "epoch": 0.07272727272727272,
+      "grad_norm": 0.09205172210931778,
+      "learning_rate": 5e-05,
+      "loss": 1.2196,
       "step": 1
     },
     {
-      "epoch": 0.7619047619047619,
-      "grad_norm": 0.5138827336171068,
-      "learning_rate": 0.0,
-      "loss": 1.506,
+      "epoch": 0.14545454545454545,
+      "grad_norm": 0.08487003296613693,
+      "learning_rate": 0.0001,
+      "loss": 1.1877,
       "step": 2
     },
     {
-      "epoch": 0.7619047619047619,
-      "step": 2,
-      "total_flos": 2446118092800.0,
-      "train_loss": 1.5153563022613525,
-      "train_runtime": 280.4756,
-      "train_samples_per_second": 0.578,
-      "train_steps_per_second": 0.007
+      "epoch": 0.21818181818181817,
+      "grad_norm": 0.6594376564025879,
+      "learning_rate": 9.797464868072488e-05,
+      "loss": 1.3252,
+      "step": 3
+    },
+    {
+      "epoch": 0.2909090909090909,
+      "grad_norm": 1.960582971572876,
+      "learning_rate": 9.206267664155907e-05,
+      "loss": 2.0162,
+      "step": 4
+    },
+    {
+      "epoch": 0.36363636363636365,
+      "grad_norm": 1.3530547618865967,
+      "learning_rate": 8.274303669726426e-05,
+      "loss": 1.5829,
+      "step": 5
+    },
+    {
+      "epoch": 0.43636363636363634,
+      "grad_norm": 1.9950076341629028,
+      "learning_rate": 7.077075065009433e-05,
+      "loss": 1.4982,
+      "step": 6
+    },
+    {
+      "epoch": 0.509090909090909,
+      "grad_norm": 0.38373228907585144,
+      "learning_rate": 5.7115741913664264e-05,
+      "loss": 1.4252,
+      "step": 7
+    },
+    {
+      "epoch": 0.5818181818181818,
+      "grad_norm": 0.14959090948104858,
+      "learning_rate": 4.288425808633575e-05,
+      "loss": 1.1972,
+      "step": 8
+    },
+    {
+      "epoch": 0.6545454545454545,
+      "grad_norm": 0.10391217470169067,
+      "learning_rate": 2.9229249349905684e-05,
+      "loss": 1.2054,
+      "step": 9
+    },
+    {
+      "epoch": 0.7272727272727273,
+      "grad_norm": 0.0690321996808052,
+      "learning_rate": 1.725696330273575e-05,
+      "loss": 1.168,
+      "step": 10
+    },
+    {
+      "epoch": 0.7272727272727273,
+      "eval_loss": 1.1994872093200684,
+      "eval_runtime": 40.5127,
+      "eval_samples_per_second": 2.419,
+      "eval_steps_per_second": 0.321,
+      "step": 10
+    },
+    {
+      "epoch": 0.8,
+      "grad_norm": 0.0676456168293953,
+      "learning_rate": 7.937323358440935e-06,
+      "loss": 1.1932,
+      "step": 11
+    },
+    {
+      "epoch": 0.8727272727272727,
+      "grad_norm": 0.061955519020557404,
+      "learning_rate": 2.0253513192751373e-06,
+      "loss": 1.2116,
+      "step": 12
+    },
+    {
+      "epoch": 0.9454545454545454,
+      "grad_norm": 0.059567924588918686,
+      "learning_rate": 0.0,
+      "loss": 1.1849,
+      "step": 13
+    },
+    {
+      "epoch": 0.9454545454545454,
+      "step": 13,
+      "total_flos": 45330292801536.0,
+      "train_loss": 1.3396454407618597,
+      "train_runtime": 2207.3438,
+      "train_samples_per_second": 0.398,
+      "train_steps_per_second": 0.006
     }
   ],
   "logging_steps": 1,
-  "max_steps": 2,
+  "max_steps": 13,
   "num_input_tokens_seen": 0,
   "num_train_epochs": 1,
-  "save_steps": 5,
+  "save_steps": 100,
   "stateful_callbacks": {
     "TrainerControl": {
       "args": {
@@ -49,7 +134,7 @@
       "attributes": {}
     }
   },
-  "total_flos": 2446118092800.0,
+  "total_flos": 45330292801536.0,
   "train_batch_size": 1,
   "trial_name": null,
   "trial_params": null