初始化项目,由ModelHub XC社区提供模型

Model: waleko/Qwen3-8B-SFT-envbench_qwen-all
Source: Original Platform
This commit is contained in:
ModelHub XC
2026-05-28 15:20:26 +08:00
commit 33c7070d28
48 changed files with 312454 additions and 0 deletions

9
train_results.json Normal file
View File

@@ -0,0 +1,9 @@
{
"epoch": 5.0,
"num_input_tokens_seen": 36600520,
"total_flos": 1.6620454705385964e+18,
"train_loss": 0.11312562568641421,
"train_runtime": 4017.5693,
"train_samples_per_second": 1.325,
"train_steps_per_second": 0.083
}