Upload V2: stackformer GPT-2 + sparse cross-attention vision model (128 visual tokens)

4666d47

verified

gurumurthy3 commited on about 12 hours ago