Hello,
Many thanks for putting together the SmolVLM work. It's great to have this great model and associated report. Would it be possible for you to release intermediate checkpoints during the training? It would be great to have access to them for specific mechanistic interpretability work on VLMs.
Many thanks,
Alessandro
Hello,
Many thanks for putting together the SmolVLM work. It's great to have this great model and associated report. Would it be possible for you to release intermediate checkpoints during the training? It would be great to have access to them for specific mechanistic interpretability work on VLMs.
Many thanks,
Alessandro