laion/nemotron-100000-opt100k__Qwen3-8B
1
0
Fork 0
Code Issues Pull Requests Actions Projects Releases Wiki Activity

初始化项目，由ModelHub XC社区提供模型

Browse Source 
Model: laion/nemotron-100000-opt100k__Qwen3-8B
Source: Original Platform
This commit is contained in:
ModelHub XC
2026-06-01 14:07:18 +08:00
commit dae8e6e845
23 changed files with 165181 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

16
all_results.json Normal file
View File

@@ -0,0 +1,16 @@
{
"achieved_tflops_per_gpu": 627863.6143054395,
"achieved_tflops_per_gpu_theoretical": 17392837.99623983,
"epoch": 5.0,
"loss_nan_ranks": 0,
"loss_rank_avg": 0.15339435636997223,
"mfu_percent": 44371.9868767095,
"mfu_percent_theoretical": 1229175.830122956,
"total_flos": 2.038899186550912e+19,
"train_loss": 0.0,
"train_runtime": 1.0148,
"train_samples_per_second": 492687.515,
"train_steps_per_second": 5133.804,
"valid_targets_mean": 9953.1,
"valid_targets_min": 2228
}