from langchain_huggingface.llms import HuggingFaceEndpoint
print("Initializing model...")
llm = HuggingFaceEndpoint(
model="nvidia/NVIDIA-Nemotron-Nano-12B-v2",
max_new_tokens=256,
temperature=0.7
)
output=llm.invoke("Can you tell me the capital of France")
print(output)
I have tried many models but got the same error.
ValueError Traceback (most recent call last)
Cell In[48], line 15
8 print("Initializing model...")
9 llm = HuggingFaceEndpoint(
10 model="nvidia/NVIDIA-Nemotron-Nano-12B-v2",
11 max_new_tokens=256,
12 temperature=0.7
13 )
---> 15 output=llm.invoke("Can you tell me the capital of Russia")
16 print(output)
File ~/Codes/VS CODE/Learning_Hugging_Face/venv/lib/python3.11/site-packages/langchain_core/language_models/llms.py:370, in BaseLLM.invoke(self, input, config, stop, **kwargs)
359 @OverRide
360 def invoke(
361 self,
(...) 366 **kwargs: Any,
367 ) -> str:
368 config = ensure_config(config)
369 return (
--> 370 self.generate_prompt(
371 [self._convert_input(input)],
372 stop=stop,
373 callbacks=config.get("callbacks"),
374 tags=config.get("tags"),
375 metadata=config.get("metadata"),
376 run_name=config.get("run_name"),
377 run_id=config.pop("run_id", None),
378 **kwargs,
379 )
380 .generations[0][0]
381 .text
382 )
File ~/Codes/VS CODE/Learning_Hugging_Face/venv/lib/python3.11/site-packages/langchain_core/language_models/llms.py:769, in BaseLLM.generate_prompt(self, prompts, stop, callbacks, **kwargs)
760 @OverRide
761 def generate_prompt(
762 self,
(...) 766 **kwargs: Any,
767 ) -> LLMResult:
768 prompt_strings = [p.to_string() for p in prompts]
--> 769 return self.generate(prompt_strings, stop=stop, callbacks=callbacks, **kwargs)
File ~/Codes/VS CODE/Learning_Hugging_Face/venv/lib/python3.11/site-packages/langchain_core/language_models/llms.py:986, in BaseLLM.generate(self, prompts, stop, callbacks, tags, metadata, run_name, run_id, **kwargs)
967 if (self.cache is None and get_llm_cache() is None) or self.cache is False:
968 run_managers = [
969 callback_manager.on_llm_start(
970 self._serialized,
(...) 984 )
985 ]
--> 986 return self._generate_helper(
987 prompts,
988 stop,
989 run_managers,
990 new_arg_supported=bool(new_arg_supported),
991 **kwargs,
992 )
993 if len(missing_prompts) > 0:
994 run_managers = [
995 callback_managers[idx].on_llm_start(
996 self._serialized,
(...) 1003 for idx in missing_prompt_idxs
1004 ]
File ~/Codes/VS CODE/Learning_Hugging_Face/venv/lib/python3.11/site-packages/langchain_core/language_models/llms.py:795, in BaseLLM._generate_helper(self, prompts, stop, run_managers, new_arg_supported, **kwargs)
784 def _generate_helper(
785 self,
786 prompts: list[str],
(...) 791 **kwargs: Any,
792 ) -> LLMResult:
793 try:
794 output = (
--> 795 self._generate(
796 prompts,
797 stop=stop,
798 # TODO: support multiple run managers
799 run_manager=run_managers[0] if run_managers else None,
800 **kwargs,
801 )
802 if new_arg_supported
803 else self._generate(prompts, stop=stop)
804 )
805 except BaseException as e:
806 for run_manager in run_managers:
File ~/Codes/VS CODE/Learning_Hugging_Face/venv/lib/python3.11/site-packages/langchain_core/language_models/llms.py:1472, in LLM._generate(self, prompts, stop, run_manager, **kwargs)
1469 new_arg_supported = inspect.signature(self._call).parameters.get("run_manager")
1470 for prompt in prompts:
1471 text = (
-> 1472 self._call(prompt, stop=stop, run_manager=run_manager, **kwargs)
1473 if new_arg_supported
1474 else self._call(prompt, stop=stop, **kwargs)
1475 )
1476 generations.append([Generation(text=text)])
1477 return LLMResult(generations=generations)
File ~/Codes/VS CODE/Learning_Hugging_Face/venv/lib/python3.11/site-packages/langchain_huggingface/llms/huggingface_endpoint.py:317, in HuggingFaceEndpoint._call(self, prompt, stop, run_manager, **kwargs)
314 completion += chunk.text
315 return completion
--> 317 response_text = self.client.text_generation(
318 prompt=prompt,
319 model=self.model,
320 **invocation_params,
321 )
323 # Maybe the generation has stopped at one of the stop sequences:
324 # then we remove this stop sequence from the end of the generated text
325 for stop_seq in invocation_params["stop"]:
File ~/Codes/VS CODE/Learning_Hugging_Face/venv/lib/python3.11/site-packages/huggingface_hub/inference/_client.py:2357, in InferenceClient.text_generation(self, prompt, details, stream, model, adapter_id, best_of, decoder_input_details, do_sample, frequency_penalty, grammar, max_new_tokens, repetition_penalty, return_full_text, seed, stop, stop_sequences, temperature, top_k, top_n_tokens, top_p, truncate, typical_p, watermark)
2355 model_id = model or self.model
2356 provider_helper = get_provider_helper(self.provider, task="text-generation", model=model_id)
-> 2357 request_parameters = provider_helper.prepare_request(
2358 inputs=prompt,
2359 parameters=parameters,
2360 extra_payload={"stream": stream},
2361 headers=self.headers,
2362 model=model_id,
2363 api_key=self.token,
2364 )
2366 # Handle errors separately for more precise error messages
2367 try:
File ~/Codes/VS CODE/Learning_Hugging_Face/venv/lib/python3.11/site-packages/huggingface_hub/inference/_providers/_common.py:93, in TaskProviderHelper.prepare_request(self, inputs, parameters, headers, model, api_key, extra_payload)
90 api_key = self._prepare_api_key(api_key)
92 # mapped model from HF model ID
---> 93 provider_mapping_info = self._prepare_mapping_info(model)
95 # default HF headers + user headers (to customize in subclasses)
96 headers = self._prepare_headers(headers, api_key)
File ~/Codes/VS CODE/Learning_Hugging_Face/venv/lib/python3.11/site-packages/huggingface_hub/inference/_providers/_common.py:171, in TaskProviderHelper._prepare_mapping_info(self, model)
168 raise ValueError(f"Model {model} is not supported by provider {self.provider}.")
170 if provider_mapping.task != self.task:
--> 171 raise ValueError(
172 f"Model {model} is not supported for task {self.task} and provider {self.provider}. "
173 f"Supported task: {provider_mapping.task}."
174 )
176 if provider_mapping.status == "staging":
177 logger.warning(
178 f"Model {model} is in staging mode for provider {self.provider}. Meant for test purposes only."
179 )
ValueError: Model nvidia/NVIDIA-Nemotron-Nano-12B-v2 is not supported for task text-generation and provider nebius. Supported task: conversational.
from langchain_huggingface.llms import HuggingFaceEndpoint
print("Initializing model...")
llm = HuggingFaceEndpoint(
model="nvidia/NVIDIA-Nemotron-Nano-12B-v2",
max_new_tokens=256,
temperature=0.7
)
output=llm.invoke("Can you tell me the capital of France")
print(output)
I have tried many models but got the same error.
ValueError Traceback (most recent call last)
Cell In[48], line 15
8 print("Initializing model...")
9 llm = HuggingFaceEndpoint(
10 model="nvidia/NVIDIA-Nemotron-Nano-12B-v2",
11 max_new_tokens=256,
12 temperature=0.7
13 )
---> 15 output=llm.invoke("Can you tell me the capital of Russia")
16 print(output)
File ~/Codes/VS CODE/Learning_Hugging_Face/venv/lib/python3.11/site-packages/langchain_core/language_models/llms.py:370, in BaseLLM.invoke(self, input, config, stop, **kwargs)
359 @OverRide
360 def invoke(
361 self,
(...) 366 **kwargs: Any,
367 ) -> str:
368 config = ensure_config(config)
369 return (
--> 370 self.generate_prompt(
371 [self._convert_input(input)],
372 stop=stop,
373 callbacks=config.get("callbacks"),
374 tags=config.get("tags"),
375 metadata=config.get("metadata"),
376 run_name=config.get("run_name"),
377 run_id=config.pop("run_id", None),
378 **kwargs,
379 )
380 .generations[0][0]
381 .text
382 )
File ~/Codes/VS CODE/Learning_Hugging_Face/venv/lib/python3.11/site-packages/langchain_core/language_models/llms.py:769, in BaseLLM.generate_prompt(self, prompts, stop, callbacks, **kwargs)
760 @OverRide
761 def generate_prompt(
762 self,
(...) 766 **kwargs: Any,
767 ) -> LLMResult:
768 prompt_strings = [p.to_string() for p in prompts]
--> 769 return self.generate(prompt_strings, stop=stop, callbacks=callbacks, **kwargs)
File ~/Codes/VS CODE/Learning_Hugging_Face/venv/lib/python3.11/site-packages/langchain_core/language_models/llms.py:986, in BaseLLM.generate(self, prompts, stop, callbacks, tags, metadata, run_name, run_id, **kwargs)
967 if (self.cache is None and get_llm_cache() is None) or self.cache is False:
968 run_managers = [
969 callback_manager.on_llm_start(
970 self._serialized,
(...) 984 )
985 ]
--> 986 return self._generate_helper(
987 prompts,
988 stop,
989 run_managers,
990 new_arg_supported=bool(new_arg_supported),
991 **kwargs,
992 )
993 if len(missing_prompts) > 0:
994 run_managers = [
995 callback_managers[idx].on_llm_start(
996 self._serialized,
(...) 1003 for idx in missing_prompt_idxs
1004 ]
File ~/Codes/VS CODE/Learning_Hugging_Face/venv/lib/python3.11/site-packages/langchain_core/language_models/llms.py:795, in BaseLLM._generate_helper(self, prompts, stop, run_managers, new_arg_supported, **kwargs)
784 def _generate_helper(
785 self,
786 prompts: list[str],
(...) 791 **kwargs: Any,
792 ) -> LLMResult:
793 try:
794 output = (
--> 795 self._generate(
796 prompts,
797 stop=stop,
798 # TODO: support multiple run managers
799 run_manager=run_managers[0] if run_managers else None,
800 **kwargs,
801 )
802 if new_arg_supported
803 else self._generate(prompts, stop=stop)
804 )
805 except BaseException as e:
806 for run_manager in run_managers:
File ~/Codes/VS CODE/Learning_Hugging_Face/venv/lib/python3.11/site-packages/langchain_core/language_models/llms.py:1472, in LLM._generate(self, prompts, stop, run_manager, **kwargs)
1469 new_arg_supported = inspect.signature(self._call).parameters.get("run_manager")
1470 for prompt in prompts:
1471 text = (
-> 1472 self._call(prompt, stop=stop, run_manager=run_manager, **kwargs)
1473 if new_arg_supported
1474 else self._call(prompt, stop=stop, **kwargs)
1475 )
1476 generations.append([Generation(text=text)])
1477 return LLMResult(generations=generations)
File ~/Codes/VS CODE/Learning_Hugging_Face/venv/lib/python3.11/site-packages/langchain_huggingface/llms/huggingface_endpoint.py:317, in HuggingFaceEndpoint._call(self, prompt, stop, run_manager, **kwargs)
314 completion += chunk.text
315 return completion
--> 317 response_text = self.client.text_generation(
318 prompt=prompt,
319 model=self.model,
320 **invocation_params,
321 )
323 # Maybe the generation has stopped at one of the stop sequences:
324 # then we remove this stop sequence from the end of the generated text
325 for stop_seq in invocation_params["stop"]:
File ~/Codes/VS CODE/Learning_Hugging_Face/venv/lib/python3.11/site-packages/huggingface_hub/inference/_client.py:2357, in InferenceClient.text_generation(self, prompt, details, stream, model, adapter_id, best_of, decoder_input_details, do_sample, frequency_penalty, grammar, max_new_tokens, repetition_penalty, return_full_text, seed, stop, stop_sequences, temperature, top_k, top_n_tokens, top_p, truncate, typical_p, watermark)
2355 model_id = model or self.model
2356 provider_helper = get_provider_helper(self.provider, task="text-generation", model=model_id)
-> 2357 request_parameters = provider_helper.prepare_request(
2358 inputs=prompt,
2359 parameters=parameters,
2360 extra_payload={"stream": stream},
2361 headers=self.headers,
2362 model=model_id,
2363 api_key=self.token,
2364 )
2366 # Handle errors separately for more precise error messages
2367 try:
File ~/Codes/VS CODE/Learning_Hugging_Face/venv/lib/python3.11/site-packages/huggingface_hub/inference/_providers/_common.py:93, in TaskProviderHelper.prepare_request(self, inputs, parameters, headers, model, api_key, extra_payload)
90 api_key = self._prepare_api_key(api_key)
92 # mapped model from HF model ID
---> 93 provider_mapping_info = self._prepare_mapping_info(model)
95 # default HF headers + user headers (to customize in subclasses)
96 headers = self._prepare_headers(headers, api_key)
File ~/Codes/VS CODE/Learning_Hugging_Face/venv/lib/python3.11/site-packages/huggingface_hub/inference/_providers/_common.py:171, in TaskProviderHelper._prepare_mapping_info(self, model)
168 raise ValueError(f"Model {model} is not supported by provider {self.provider}.")
170 if provider_mapping.task != self.task:
--> 171 raise ValueError(
172 f"Model {model} is not supported for task {self.task} and provider {self.provider}. "
173 f"Supported task: {provider_mapping.task}."
174 )
176 if provider_mapping.status == "staging":
177 logger.warning(
178 f"Model {model} is in staging mode for provider {self.provider}. Meant for test purposes only."
179 )
ValueError: Model nvidia/NVIDIA-Nemotron-Nano-12B-v2 is not supported for task text-generation and provider nebius. Supported task: conversational.