passes_n_ocr_s2_d60m
d60m (68.9M parameters) on plus_all v2 subset + pilot OCR text (39M tokens), trained on RTX 3060 12GB (Home GPU). Started 10 Oct 2026, ended 11 Oct 2026.
Hypothesis
The N-ocr result repeats with a second seed.
What we learned
Confirms N-ocr: ex-Gita 0.6637, 0.38% behind a fourth pass (clean_v1 +0.34%). Two-seed mean +0.27% vs Passes 4 and 1.3% behind fresh clean text, so OCR text stays out of the corpus for now.
Scores
Bits per byte, lower is better; change against the parent run, Passes N-ocr.
ex-Gītā (headline)
0.6637
bits per byte
+0.2%ex-Gītā, clean_v1
0.6854
bits per byte
+0.2%Pooled, all five sets
0.6563
bits per byte
+0.2%Validation split
0.6524
bits per byte
+0.3%| Test set | Passes N-ocr s2 | Passes N-ocr (parent) | Change |
|---|---|---|---|
| DCS gold (classical) | 0.6736 | 0.6711 | +0.4% |
| Bhagavad-gītā (memorisation) | 0.4606 | 0.46 | +0.1% |
| Out of domain | 0.6678 | 0.6658 | +0.3% |
| Prose | 0.6301 | 0.6312 | −0.2% |
| Vedic (Ṛgveda) | 0.9056 | 0.8937 | +1.3% |
Curves
Training loss
Cross-entropy per token, by step. The first few percent of the run, far higher, run off the top; hover or the table has every value.
Held-out bits per byte
The validation split, evaluated during training, by step. Lower is better.
Throughput
Tokens per second, by step.
Model
- Preset
- d60m
- Parameters
- 68,924,160
- Outside embeddings
- 56,636,160
- Layers · heads · width
- 8 · 12 · 768
- Tokenizer
- SLP1 unigram 8k
Data
- Slice
- plus_all v2 subset + pilot OCR text (39M tokens)
- Words
- —
- Training tokens
- 194,623,344
- Passes
- 3.19 passes
- Tokens seen
- 621,084,672
Compute
- GPU
- RTX 3060 12GB
- Where
- Home GPU
- Steps
- 12,636 / 12,636
- GPU hours
- 3.83
- Cost
- —
- Spot restarts
- —
Lineage
homed60m