Datasets:

Nanthasit
/

sakthai-kaggle-notebooks

Tasks:

Text Generation

Languages:

Size:

Tags:

training-scripts

dataset:Nanthasit/sakthai-kaggle-notebooks

License:

Dataset card Files Files and versions

sakthai-kaggle-notebooks

Ctrl+K

Ctrl+K

1 contributor

History: 48 commits

Nanthasit's picture

Upload scripts/augment-fill-gaps.py with huggingface_hub

5c4605e verified 2 months ago

scripts
Upload scripts/augment-fill-gaps.py with huggingface_hub 2 months ago
.gitattributes

2.5 kB
initial commit 3 months ago
README.md

11.2 kB
Upload README.md with huggingface_hub 2 months ago
deploy-endpoint.py

3.22 kB
Endpoint deployment script for merged model 3 months ago
eval-bfcl-0.5b.py

4 kB
Add BFCL-style eval on the held-out combined-v6 test split 2 months ago
job-0.5b-exp.py

13.2 kB
exp: SAK_DUP_CAP + fixed NanGuard (no-labels reforward); sync-tested 2 months ago
job-0.5b-hfjobs.py

8.01 kB
0.5B job: 2 epochs, no grad-checkpoint (faster), max_len 1536, checkpoint-to-hub safety 2 months ago
job-0.5b-nanguard.py

13.5 kB
NanGuard: zero grads BEFORE diagnostics; dump wrapped in try/except (first run confirmed rows have 6-77 unmasked labels at the poison step) 2 months ago
job-0.5b-nanhunt.py

12 kB
nanhunt: abort at first non-finite gradient, dump the triggering batch (v3 disproved the seed theory) 2 months ago
job-0.5b-v7.py

10.4 kB
H200: micro-batch 32 x accum 2 (batch 64 OOMs without checkpointing) 2 months ago
sakthai-7b-engine.ipynb

24.6 kB
Add 7B training notebook with validation + benchmark 3 months ago
sakthai-engine.ipynb

10.7 kB
SakThai Engine - Qwen2.5-1.5B LoRA fine-tune notebook 3 months ago
train-sakthai-0.5b-v2.py

6.91 kB
Add improved 0.5B fine-tune script (MLP LoRA targets, rsLoRA, completion-only loss, v6+irrelevance) 2 months ago
validate.py

5.06 kB
Fix validator bench path 2 months ago
validate_exp.py

3.81 kB
Add 0.5B experiment script (masking / full-FT modes); eval moves to the bench repo runner 2 months ago