End of training

README.md

@@ -4,6 +4,7 @@ license: apache-2.0
 base_model: secmlr/DS-Noisy_DS-Clean_QWQ-Noisy_QWQ-Clean_Qwen2.5-7B-Instruct_full_sft_1e-5
 tags:
 - llama-factory
+- full
 - generated_from_trainer
 model-index:
 - name: ruizhe_simplier_dsNsy32kCln32k_QwQNsy32kCln32k_DSNsyCln_QWQNsyCln_Qwen7B_summarized_sft
@@ -15,7 +16,7 @@ should probably proofread and complete it, then remove this comment. -->
 
 # ruizhe_simplier_dsNsy32kCln32k_QwQNsy32kCln32k_DSNsyCln_QWQNsyCln_Qwen7B_summarized_sft
 
-This model is a fine-tuned version of [secmlr/DS-Noisy_DS-Clean_QWQ-Noisy_QWQ-Clean_Qwen2.5-7B-Instruct_full_sft_1e-5](https://huggingface.co/secmlr/DS-Noisy_DS-Clean_QWQ-Noisy_QWQ-Clean_Qwen2.5-7B-Instruct_full_sft_1e-5) on the None dataset.
+This model is a fine-tuned version of [secmlr/DS-Noisy_DS-Clean_QWQ-Noisy_QWQ-Clean_Qwen2.5-7B-Instruct_full_sft_1e-5](https://huggingface.co/secmlr/DS-Noisy_DS-Clean_QWQ-Noisy_QWQ-Clean_Qwen2.5-7B-Instruct_full_sft_1e-5) on the ruizhe_simplier_reasoning_ds_clean_32k, the ruizhe_simplier_reasoning_ds_noisy_32k, the ruizhe_simplier_reasoning_qwq_clean_32k and the ruizhe_simplier_reasoning_qwq_noisy_small_32k datasets.
 
 ## Model description
 

all_results.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 2.993417721518987,
+    "total_flos": 2.6892487337023898e+17,
+    "train_loss": 0.5392509524643482,
+    "train_runtime": 11583.2622,
+    "train_samples_per_second": 4.092,
+    "train_steps_per_second": 0.085
+}

train_results.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 2.993417721518987,
+    "total_flos": 2.6892487337023898e+17,
+    "train_loss": 0.5392509524643482,
+    "train_runtime": 11583.2622,
+    "train_samples_per_second": 4.092,
+    "train_steps_per_second": 0.085
+}