Downloads · 30 days
13
3% of all-time downloads
Qwen3.5-4B-DocVQA-ReAct-Agentic-ESOpt
zz1358m/Qwen3.5-4B-DocVQA-ReAct-Agentic-ESOpt
Qwen3.5-4B-DocVQA-ReAct-Agentic-ESOpt is a text generation model from zz1358m. Use it when you need the model to write or continue text. It is set up for transformers.
Hugging Face checkpoint reconstructed from the Agentic ESOpt DocVQA run.
All-time downloads
453
Public
Parameters
4.2B
8.4 GB on disk
Likes
0
Public
Repo makeup
Click a slice to open those files.
.safetensors8.4 GB · 100%
Parameter types
How the weights are stored.
BF164.2B · 100%
Try a prompt
Base models
Model card
From the Hugging Face model README
Qwen3.5-4B-DocVQA-ReAct-Agentic-ESOpt
Hugging Face checkpoint reconstructed from the Agentic ESOpt DocVQA run.
- Base runtime checkpoint:
Qwen3.5-4B-text - ES scope: full model, 330 tensors / 4,205,751,296 parameters
- Applied updates: 40
- Included generations: 0 through 39
- Update rule: stored ES seeds and z-score-normalized weights, alpha 0.0005
- Weight format: two indexed safetensors shards
- Verified loader:
transformers.AutoModelForCausalLM
replay_metadata.json records the local lineage and replayed_history.json
contains the exact 40 update records used to reconstruct this checkpoint.
from transformers import AutoModelForCausalLM, AutoTokenizer
path = "zz1358m/Qwen3.5-4B-DocVQA-ReAct-Agentic-ESOpt"
tokenizer = AutoTokenizer.from_pretrained(path)
model = AutoModelForCausalLM.from_pretrained(path, dtype="auto")
- Task
- Text Generation
- Library
- transformers
- Type
- qwen3_next
- Created
- Aug 9, 2026
- Updated
- Aug 11, 2026