Downloads · 30 days
9
24% of all-time downloads
emorf1003/bam-vl-qwen35-27b
bam-vl-qwen35-27b is a image-text-to-text model from emorf1003. Use it for the image-text-to-text task on the model card, and read the license before you ship it in a product. It is set up for transformers. The card lists the license as other.
LoRA fine-tune of Qwen/Qwen3.5-27B for video question answering, exported from a BAM-VL training checkpoint (step2023, globalstep 2023).
Downloads · 30 days
9
24% of all-time downloads
All-time downloads
38
Public
Parameters
27.4B
54.7 GB on disk
Likes
0
Public
Click a slice to open those files.
.safetensors54.7 GB · 100%
From the Hugging Face model README
LoRA fine-tune of Qwen/Qwen3.5-27B for video question answering, exported from a BAM-VL
training checkpoint (step_2023, global_step 2023).
AutoModelForImageTextToText.bam_adapters.pt — the BAM side-channel adapters (facial / pose / audio) and CLS heads,
with the hyperparameters needed to rebuild them.The merged backbone alone does not reproduce training behaviour. During training the
model pools the penultimate hidden state, computes a delta from the side-channel adapters,
and adds it to the final hidden state before lm_head. A plain generate() call skips
this entirely. Faithful inference requires bam_adapters.pt together with the BAMVLQA
wrapper code and the corresponding facial/pose/audio feature extractors.
bam_only — frozen backbone, side-channel adapters only.bam_and_full_model — LoRA (r=32, alpha=64) plus adapters.