OpenGVLab/VideoChat-R1_7B_caption

Partial fine-tune vs Qwen/Qwen2-VL-7B-Instruct. Partial change: 85% of the weights changed (attention, lm_head, mlp, norms), in 28 of 32 layers.

Tensors changed: Embeddings 0/2, Norms 52/187, MLP 84/216, Attention 194/324, lm_head 1/1

1 derivatives checked

• mradermacher/VideoChat-R1_7B_caption-GGUF: Same weights, converted (fewer layers)

On the Hub · Downloads on Model Pulse