Downloads · 30 days
0
qwen3-4b-ocr-think-stage2-merged
modrill/qwen3-4b-ocr-think-stage2-merged
qwen3-4b-ocr-think-stage2-merged is a machine learning model from modrill. Use it for the machine learning task on the model card, and read the license before you ship it in a product.
Qwen3-4B OCR LoRA merged model (stage2, untiedinputonly repatch).
Access
Public
Updated Jul 3, 2026
Repo size
8.8 GB
Likes
0
Public
Repo makeup
Click a slice to open those files.
.tar8.8 GB · 100%
At a glance
- Access
- Public
- Created
- Jul 3, 2026
- Updated
- Jul 3, 2026
- SHA
- 25a27bce
Model card
From the Hugging Face model README
Qwen3-4B OCR Think (stage2 merged)
Qwen3-4B OCR LoRA merged model (stage2, untied_input_only repatch).
- Base: stoprows-baked + attn/MLP LoRA merged
tie_word_embeddings=false(independent embed / lm_head)- Package:
qwen3-think-stage2-merged.tar(8.3G)
Load
tar -xf qwen3-think-stage2-merged.tar
# then
from transformers import AutoModelForCausalLM, AutoTokenizer
m = AutoModelForCausalLM.from_pretrained("./qwen3-*-stage2-merged", trust_remote_code=True)
See LOADING_NOTE.md inside the tarball for transfer caveats.
- Created
- Jul 3, 2026
- Updated
- Jul 3, 2026