INNER CODE UNIT · Python

bot_streaming_T2T

FoundationVision/Liquid · evaluation/app.py:279

def bot_streaming_T2T(message, history,temperature):
    print(message)
    global stop_flag
    stop_flag = True
    time.sleep(0.2)
    stop_flag = False
    torch.cuda.empty_cache()
    qs = message 
    conv = conv_templates['gemma'].copy()
    conv.append_message(conv.roles[0], qs)
    conv.append_message(conv.roles[1], None)
    prompt = conv.get_prompt()
 
    print(prompt)
    with torch.no_grad():
        inputs = tokenizer([prompt], return_tensors="pt").to('cuda')
        streamer = TextIteratorStreamer(tokenizer, **{"skip_special_tokens": False, "skip_prompt": True})

View source record →