初始化项目,由ModelHub XC社区提供模型

Model: Neelectric/Llama-3.1-8B-Instruct_SFT_mathsp_ewc_v00.03
Source: Original Platform
This commit is contained in:
ModelHub XC
2026-05-10 05:37:33 +08:00
commit e4914c4917
17 changed files with 285717 additions and 0 deletions

11
train_results.json Normal file
View File

@@ -0,0 +1,11 @@
{
"ewc_loss": 0.224609375,
"ewc_loss_diag": 2.4199485778808594e-05,
"ewc_loss_parallel": 0.0002002716064453125,
"total_flos": 5.62815163329864e+19,
"train_loss": 0.5611200594813073,
"train_runtime": 47572.7611,
"train_samples": 125770,
"train_samples_per_second": 7.931,
"train_steps_per_second": 0.496
}