randhir302
/

MoEVL-Tiny-0.5B

Image-Text-to-Text

Mixture of Experts

vision-language

Model card Files Files and versions

MoEVL-Tiny-0.5B

1.77 GB

Ctrl+K

Ctrl+K

1 contributor

History: 19 commits

randhir302's picture

CRITICAL FIX: Corrected MoE structure in HF deploy code and resized token embeddings.

ab99517 verified 11 days ago

tokenizer
Upload self-contained merged model 15 days ago
vision_encoder
Upload self-contained merged model 15 days ago
.gitattributes

1.58 kB
Upload self-contained merged model 15 days ago
README.md

5.79 kB
Update README.md 11 days ago
config.json

1.3 kB
CRITICAL FIX: Corrected MoE structure in HF deploy code and resized token embeddings. 11 days ago
generation_config.json

241 Bytes
Upload self-contained merged model 15 days ago
model.safetensors

902 MB
xet

CRITICAL FIX: Corrected MoE structure in HF deploy code and resized token embeddings. 11 days ago
modeling_moevl_tiny.py

7.62 kB
CRITICAL FIX: Corrected MoE structure in HF deploy code and resized token embeddings. 11 days ago
projector.pt
Detected Pickle imports (3)
- "torch.BFloat16Storage",
- "collections.OrderedDict",
- "torch._utils._rebuild_tensor_v2"
What is a pickle import?
3.67 MB
xet

Clean full‑precision model (PixMo checkpoint, better grounding) 11 days ago