End of training

README.md

@@ -4,6 +4,7 @@ license: apache-2.0
 base_model: Qwen/Qwen3-8B
 tags:
 - llama-factory
+- full
 - generated_from_trainer
 model-index:
 - name: a1_math_formulas
@@ -15,7 +16,7 @@ should probably proofread and complete it, then remove this comment. -->
 
 # a1_math_formulas
 
-This model is a fine-tuned version of [Qwen/Qwen3-8B](https://huggingface.co/Qwen/Qwen3-8B) on an unknown dataset.
+This model is a fine-tuned version of [Qwen/Qwen3-8B](https://huggingface.co/Qwen/Qwen3-8B) on the mlfoundations-dev/a1_math_formulas dataset.
 
 ## Model description
 

all_results.json

@@ -1,12 +1,16 @@
 {
+    "achieved_tflops_per_gpu": 2.476802423960275,
+    "achieved_tflops_per_gpu_theoretical": 481.0824479208898,
     "epoch": 5.0,
     "loss_nan_ranks": 0,
-    "loss_rank_avg": 0.018037570640444756,
-    "total_flos": 2.10289015818027e+18,
-    "train_loss": 0.2609484052790804,
-    "train_runtime": 28006.4544,
-    "train_samples_per_second": 5.642,
-    "train_steps_per_second": 0.044,
-    "valid_targets_mean": 4350.2,
-    "valid_targets_min": 878
+    "loss_rank_avg": 0.29739144444465637,
+    "mfu_percent": 0.1750390405625636,
+    "mfu_percent_theoretical": 33.998759570380905,
+    "total_flos": 1.3775029894633226e+18,
+    "train_loss": 0.31707094306887884,
+    "train_runtime": 34760.1149,
+    "train_samples_per_second": 4.545,
+    "train_steps_per_second": 0.142,
+    "valid_targets_mean": 4348.2,
+    "valid_targets_min": 876
 }

train_results.json

@@ -1,12 +1,16 @@
 {
+    "achieved_tflops_per_gpu": 2.476802423960275,
+    "achieved_tflops_per_gpu_theoretical": 481.0824479208898,
     "epoch": 5.0,
     "loss_nan_ranks": 0,
-    "loss_rank_avg": 0.018037570640444756,
-    "total_flos": 2.10289015818027e+18,
-    "train_loss": 0.2609484052790804,
-    "train_runtime": 28006.4544,
-    "train_samples_per_second": 5.642,
-    "train_steps_per_second": 0.044,
-    "valid_targets_mean": 4350.2,
-    "valid_targets_min": 878
+    "loss_rank_avg": 0.29739144444465637,
+    "mfu_percent": 0.1750390405625636,
+    "mfu_percent_theoretical": 33.998759570380905,
+    "total_flos": 1.3775029894633226e+18,
+    "train_loss": 0.31707094306887884,
+    "train_runtime": 34760.1149,
+    "train_samples_per_second": 4.545,
+    "train_steps_per_second": 0.142,
+    "valid_targets_mean": 4348.2,
+    "valid_targets_min": 876
 }