| | import os |
| |
|
| | from huggingface_hub import HfApi |
| |
|
| | API = HfApi(token=os.environ.get("HF_TOKEN")) |
| | REPO_ID = "Limour/llama-python-streamingllm" |
| |
|
| |
|
| | def restart_space(): |
| | API.restart_space(repo_id=REPO_ID, token=os.environ.get("HF_TOKEN")) |
| |
|
| |
|
| | def init(cfg): |
| | |
| | model = cfg['model'] |
| | s_info = cfg['s_info'] |
| |
|
| | def btn_reset(_cache_path): |
| | try: |
| | with cfg['session_lock']: |
| | _tmp = model.load_session(_cache_path) |
| | print(f'load cache from {_cache_path} {_tmp}') |
| | cfg['session_active'] = True |
| | return model.venv_info |
| | except Exception as e: |
| | restart_space() |
| | raise e |
| |
|
| | def btn_stop(): |
| | cfg['btn_stop_status'] = True |
| |
|
| | cfg['btn_stop'].click( |
| | fn=btn_stop |
| | ) |
| |
|
| | cfg['btn_reset'].click( |
| | fn=btn_stop |
| | ).success( |
| | fn=btn_reset, |
| | inputs=cfg['setting_cache_path'], |
| | outputs=s_info, |
| | **cfg['btn_concurrency'] |
| | ).success( |
| | **cfg['btn_finish'] |
| | ) |
| |
|
| | cfg['btn_debug'].click( |
| | fn=lambda: model.str_detokenize(model._input_ids), |
| | outputs=cfg['vo'] |
| | ) |
| |
|