btbtyler09
/

shrew-ocr-preview

Image-Text-to-Text

granite4_vision

document-intelligence

structured-extraction

vision-language

Model card Files Files and versions

shrew-ocr-preview

Commit History

Sep 28 30 days ago

00:09
v0.4 — merged bf16, v0.4: reinforcement-learning stage (GRPO, r=256 LM LoRA, ckpt-3000) merged onto the unpublished full-epoch merge of v0.3; shipped with the sampled-first-pass recipe (temp 0.4 / presence 0.3, one un-enforced retry, then fallback); OHR-Bench paired retrieval vs v0.3 weights on the same server 394:180 (p 2e-19), failed pages 0.32% vs 1.00%; requires shrew-server >= 0.4.0

btbtyler09

Verified signature
bb82611

Sep 14

11:00
v0.3 — merged bf16, generation E6: LoRA granite_4b_dense_e6 (r=256, 1 epoch / 1,343 steps, eval_loss 0.1408) merged into the v0.2 bucket base; pinpoints unchanged; requires shrew-server >= 0.3.14 (36-value section_type contract)

btbtyler09

Verified signature
0a832e9

Aug 17

16:58
README: GGUF Versions row — full context per slot rule

btbtyler09

Verified signature
4bcf71d
11:47
README: shrew-server pin v0.2.0 -> v0.2.1

btbtyler09

Verified signature
a9b66b9

Aug 16

23:17
README: add GGUF variant to Versions table

btbtyler09

Verified signature
5d76834
22:07
README: measured decode guidance (presence_penalty 0.3 first-pass default, 0.6 retry rescue; enforcement retry-only)

btbtyler09

Verified signature
4ffd73d
22:03
README: link shrew-server reference pipeline; add --served-model-name to vLLM examples

btbtyler09

Verified signature
eb93514
21:53
README: correct penalty-mechanics wording (frequency accumulates; repetition is flat per seen token)

btbtyler09

Verified signature
bc8c224
21:53
README: technical-manual voice; add text-modality spec and penalty-retry guidance

btbtyler09

Verified signature
098e7b3
20:16
e4 generation: LM-only LoRA r=256 merged, bucket-routed tiling (2026-08-16)

btbtyler09

Verified signature
c0b1001
20:15
initial commit

btbtyler09

Verified signature
2cac452