初始化项目,由ModelHub XC社区提供模型

Model: Neelectric/Llama-3.1-8B-Instruct_SFT_mathsp_ewc_v00.04
Source: Original Platform
This commit is contained in:
ModelHub XC
2026-05-22 17:14:37 +08:00
commit 60ea35c3dc
17 changed files with 285717 additions and 0 deletions

11
train_results.json Normal file
View File

@@ -0,0 +1,11 @@
{
"ewc_loss": 0.0849609375,
"ewc_loss_diag": 3.457069396972656e-05,
"ewc_loss_parallel": 0.0005035400390625,
"total_flos": 5.62815163329864e+19,
"train_loss": 0.4737072293779522,
"train_runtime": 48395.8245,
"train_samples": 125770,
"train_samples_per_second": 7.796,
"train_steps_per_second": 0.487
}