Upload app (Copy).py

app (Copy).py

@@ -16,9 +16,6 @@ from transformers import pipeline, AutoTokenizer, Gemma3ForCausalLM
 logging.getLogger('numba').setLevel(logging.WARNING)
 limitation = os.getenv("SYSTEM") == "spaces"  # limit text and audio length in huggingface spaces
 
-# Set an environment variable export HF_HOME='/media/shiroclancy/Local Disk/tma'
-os.environ['HF_HOME'] = '/media/shiroclancy/Local Disk/tma'
-
 hps_ms = utils.get_hparams_from_file(r'config/config.json')
 
 audio_postprocess_ori = gr.Audio.postprocess
@@ -66,7 +63,7 @@ def create_tts_fn(net_g_ms, speaker_id):
                                    length_scale=length_scale)[0][0, 0].data.cpu().float().numpy()
 
         #return "Success", (22050, audio)
-        return english, (22050, audio)
+        return english, japanese, (22050, audio)
     return tts_fn
 
 def create_to_symbol_fn(hps):
@@ -140,26 +137,19 @@ messages = init_character(
     """
 )
 
-def generate_response(pipe, human_prompt, tokenizer = None):
+def generate_response(model, human_prompt, tokenizer = None):
     # Add user message
-    #device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
     messages.append({
         "role": "user",
         "content": [{"type": "text", "text": human_prompt}],
     })
-
-    # Decode
-    response = pipe(messages, max_new_tokens = 2048, temperature = 0.2)[0]['generated_text'][-1]['content']#, do_sample = False)
+    response = model.create_chat_completion(messages = messages)['choices'][0]['message']['content']
 
     # Save assistant reply to history   
     #print(response)
     english, japanese = response.split(" <split> ")
-    
-    messages.append({
-        "role": "assistant",
-        "content": [{"type": "text", "text": english}],
-    })
-    #messages.pop()
+
+    messages.pop()
     return english, japanese
 
 download_audio_js = """
@@ -183,13 +173,15 @@ download_audio_js = """
     oA.remove();
 }}
 """
-model_id = "google/gemma-3-4b-it"
-pipe = pipeline("text-generation", 
-                    model=model_id, 
-                    #tokenizer=tokenizer, 
-                    #device=device, 
-                    device_map ='auto',
-                    torch_dtype=torch.bfloat16)
+from llama_cpp import Llama
+
+llm = Llama.from_pretrained(
+	repo_id="google/gemma-3-4b-it-qat-q4_0-gguf",
+    filename='gemma-3-4b-it-q4_0.gguf',
+	local_dir='/kaggle/working/model',
+    n_ctx = 2048
+)
+
 if __name__ == '__main__':
     device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')             
     parser = argparse.ArgumentParser()
@@ -264,10 +256,11 @@ if __name__ == '__main__':
                                     nsw = gr.Slider(label="noise_scale_w", minimum=0.1, maximum=1.0, step=0.1, value=0.668, interactive=True)
                                     ls = gr.Slider(label="length_scale", minimum=0.1, maximum=2.0, step=0.1, value=1.2 if language=="Chinese" else 1, interactive=True)
                             with gr.Column():
-                                o1 = gr.Textbox(label="Output Message")
+                                o1 = gr.Textbox(label="Output Message(English)")
+                                o3 = gr.Textbox(label="Output Message(Japanese)")
                                 o2 = gr.Audio(label="Output Audio", elem_id=f"tts-audio-en-{name_en.replace(' ','')}")
                                 download = gr.Button("Download Audio")
-                            btn.click(tts_fn, inputs=[input_text, lang,  ns, nsw, ls, symbol_input], outputs=[o1, o2], api_name=f"tts-{name_en}")
+                            btn.click(tts_fn, inputs=[input_text, lang,  ns, nsw, ls, symbol_input], outputs=[o1, o3, o2], api_name=f"tts-{name_en}")
                             download.click(None, [], [], _js=download_audio_js.format(audio_id=f"en-{name_en.replace(' ', '')}"))
                             lang.change(change_lang, inputs=[lang], outputs=[ns, nsw, ls])
                             symbol_input.change(