fix web demo with new multi-lora interface

This commit is contained in:
51616 2025-08-13 06:33:34 +00:00
parent 3ae4cb57ce
commit dfee685678

View file

@ -652,6 +652,9 @@ def chat():
outputs = modulated_model.generate_with_multi_loras(
ctx_ids=ctx_ids,
ctx_attn_mask=ctx_attn_mask,
n_ctx_chunks=torch.tensor(
[len(ctx_ids)], device=ctx_ids.device
),
input_ids=model_inputs,
max_new_tokens=512,
do_sample=False,