初始化项目,由ModelHub XC社区提供模型

Model: W-61/llama-3-8b-base-sft-hh-helpful-4xh200
Source: Original Platform
This commit is contained in:
ModelHub XC
2026-04-25 14:10:50 +08:00
commit 1db7c7ce96
20 changed files with 3679 additions and 0 deletions

9
train_results.json Normal file
View File

@@ -0,0 +1,9 @@
{
"epoch": 0.9980657640232108,
"total_flos": 9.517157411769549e+16,
"train_loss": 1.5008004404777704,
"train_runtime": 885.0234,
"train_samples": 43598,
"train_samples_per_second": 18.662,
"train_steps_per_second": 0.292
}