(tokenizer, query, history, do_sample, max_length, top_k, top_p, temperature, repetition_penalty)
| 68 | |
| 69 | |
| 70 | def streamresponse_v2(tokenizer, query, history, do_sample, max_length, top_k, top_p, temperature, repetition_penalty): |
| 71 | result_generator = model.chat(tokenizer, query, history=history, generation_config=generate_config, stream=True, |
| 72 | do_sample=do_sample, max_length=max_length, top_k=top_k, temperature=temperature, |
| 73 | repetition_penalty=repetition_penalty, top_p=top_p) |
| 74 | t_resp = '' |
| 75 | while 1: |
| 76 | try: |
| 77 | char, _ = next(result_generator) |
| 78 | if char is None: |
| 79 | break |
| 80 | else: |
| 81 | t_resp += char |
| 82 | |
| 83 | yield char |
| 84 | except StopIteration: |
| 85 | break |
| 86 | |
| 87 | |
| 88 | def response_data(seqid, code, message, flag, data): |
no test coverage detected