Sponsored by Vmake Labs Video Enhancer - Vmake Labs Video Enhancer uses AI to improve video quality, restore details, reduce blur, and upscale videos to 8K.

AI Ad Library

AI Models

chrono-2021-sn38

kepton0117 / chrono-2021-sn38

Run on huggingface.co

huggingface.co

Total runs: 1.2K

24-hour runs: 0

7-day runs: 443

30-day runs: 443

Model's Last Updated: September 21 2026

text-generation

Introduction of chrono-2021-sn38

Model Details of chrono-2021-sn38

chrono-2021-sn38 — SN38 Round 12 (year 2021)

Continue-pretrained sn38-nanochrono (~2.02B) for Bittensor subnet 38, cutoff 2021-12-31 .

Pin (this repo): f62e238265e3ca4a4ed20b300475513446d9e1fc = train step-4000
Base: anacoluthe89/chrono-2020 @ 633eb832c3de8eceb0b5bba1c5fb94faa8e777c6 (UID 131)
Data + scripts: kepton0117/sn38-train-2021
Code: github.com/kepton0117/sn38
Cosine vs UID 131 at this pin: 0.9294 (copy gate needs < 0.95)

recipe/ in this repo is the train algorithm and launch scripts that produced the pin.

Algorithm

Full-parameter causal LM continue-pretrain (CPT). Not LoRA. Not instruction SFT.

loss = cross_entropy(model(tokens[:-1]), tokens[1:])

Piece	Value
Arch	`sn38-nanochrono` , 28 layers, hidden 1792, seq 2048
Update	all parameters
Replay	15% FineWeb 2017–2020
Packing	docs concatenated, cut every 2048 tokens, EOS between docs
Optimizer	AdamW, weight decay 0.1, betas (0.9, 0.95)
Schedule	cosine decay, 3% warmup
Parallel	8× H200 DDP

Recipe that produced this pin

Knob	Value
GPUs	8× H200
Micro-batch / GPU	16
Grad accum	2
Tokens / step	8 × 16 × 2048 × 2 = 524,288
LR	1e-3 (1e-4 / 3e-4 left cosine stuck at ~0.999)
Checkpoint	step-4000
Data	FineWeb last-4 2021 dumps (60%) + older FineWeb (15%) + `brimmann2/enwiki-dec2021` (25%)
Filter	drop short / nav / repetitive docs (~12.5%)
Packed tokens	2,610,770 × 2048 ≈ 5.35B

# needs sn38.architectures from the GitHub repo
LR=1e-3 bash scripts/run_train_h200x8_2021.sh

Do not use FineWeb name='default' or Wikipedia 20220301.en — those leak 2022+.

Files

Path	What
`model.safetensors`	step-4000 weights
`config.json`	nanochrono config
`recipe/`	launcher, trainer, data build, filter

Runs of kepton0117 chrono-2021-sn38 on huggingface.co

1.2K

Total runs

24-hour runs

3-day runs

443

7-day runs

443

30-day runs

More Information About chrono-2021-sn38 huggingface.co Model

More chrono-2021-sn38 license Visit here:

https://choosealicense.com/licenses/other

chrono-2021-sn38 huggingface.co

chrono-2021-sn38 huggingface.co is an AI model on huggingface.co that provides chrono-2021-sn38's model effect (), which can be used instantly with this kepton0117 chrono-2021-sn38 model. huggingface.co supports a free trial of the chrono-2021-sn38 model, and also provides paid use of the chrono-2021-sn38. Support call chrono-2021-sn38 model through api, including Node.js, Python, http.

chrono-2021-sn38 huggingface.co Url

https://huggingface.co/kepton0117/chrono-2021-sn38