Upload app.py

app.py

@@ -112,6 +112,7 @@ with gr.Blocks() as setting:
         os.mkdir("cache")
         snapshot_download(repo_id='TheBloke/CausalLM-7B-GGUF', local_dir=r'downloads',
                           allow_patterns='causallm_7b.Q5_K_M.gguf')
+        snapshot_download(repo_id='Limour/llama-python-streamingllm-cache', repo_type='dataset', local_dir=r'cache')
 
     #  ========== 加载模型 ==========
     model = StreamingLLM(model_path=setting_path.value,
@@ -194,6 +195,10 @@ with gr.Blocks() as role:
             pass
         tmp = model.save_session(setting_cache_path.value)
         print(f'save cache {tmp}')
+        # ========== 上传缓存 ==========
+        from huggingface_hub import login, CommitScheduler
+        login(token=os.environ.get("HF_TOKEN"), write_permission=True)
+        CommitScheduler(repo_id='Limour/llama-python-streamingllm-cache', repo_type='dataset', folder_path='cache')
 
 
 # ========== 流式输出函数 ==========