初始化项目,由ModelHub XC社区提供模型

Model: Neelectric/Llama-3.1-8B-Instruct_SFT_mathsp_ewc_v00.19
Source: Original Platform
This commit is contained in:
ModelHub XC
2026-09-02 17:37:16 +08:00
commit f2421e9cee
17 changed files with 285717 additions and 0 deletions

11
train_results.json Normal file
View File

@@ -0,0 +1,11 @@
{
"ewc_loss": 0.01806640625,
"ewc_loss_diag": 0.0,
"ewc_loss_parallel": 1.811981201171875e-05,
"total_flos": 5.62815163329864e+19,
"train_loss": 0.40474898667779935,
"train_runtime": 46250.2298,
"train_samples": 125770,
"train_samples_per_second": 8.158,
"train_steps_per_second": 0.51
}