File size: 375 Bytes
fd2ac66
 
 
 
 
99de728
 
8b08481
5a773e7
fd2ac66
f941585
99de728
1e9e124
4ad04d9
ab7480b
e12b22b
14d38e8
 
fd2ac66
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
numpy<2.0
torch==2.1.2
torchvision==0.16.2
safetensors>=0.4.3

# Qwen needs a recent Transformers; model card advises newest to avoid KeyError
transformers==4.51.3
accelerate==0.32.1
peft==0.17.0
huggingface_hub>=0.23,<1.0

qwen-vl-utils>=0.0.8
# --- UI: keep these in lockstep (no 0.x client with 5.x gradio) ---
gradio==4.44.1
gradio_client==1.3.0

pydantic==2.10.6

pillow