Sythentic Data Generation Notebook broken #69
Description
On recent colab T4 when I run the Synthetic Data Generation Notebook the following line throws below error:
Code:
from unsloth.dataprep import SyntheticDataKit
generator = SyntheticDataKit.from_pretrained(
# Choose any model from https://huggingface.co/unsloth
model_name = "unsloth/Llama-3.2-3B-Instruct",
max_seq_length = 2048, # Longer sequence lengths will be slower!
)
Error:
tokenizer_config.json:
54.7k/? [00:00<00:00, 6.33MB/s]
tokenizer.json: 100%
17.2M/17.2M [00:01<00:00, 12.6MB/s]
special_tokens_map.json: 100%
454/454 [00:00<00:00, 63.0kB/s]
chat_template.jinja:
3.83k/? [00:00<00:00, 419kB/s]
Unsloth: Using dtype = torch.float16 for vLLM.
Unsloth: vLLM loading unsloth/Llama-3.2-3B-Instruct with actual GPU utilization = 89.39%
Unsloth: Your GPU has CUDA compute capability 7.5 with VRAM = 14.74 GB.
Unsloth: Using conservativeness = 1.0. Chunked prefill tokens = 2048. Num Sequences = 192.
Unsloth: vLLM's KV Cache can use up to 7.19 GB. Also swap space = 4 GB.
vLLM STDOUT: INFO 07-12 07:47:36 [__init__.py:239] Automatically detected platform cuda.
vLLM STDOUT: INFO 07-12 07:47:41 [api_server.py:1043] vLLM API server version 0.8.5.post1
vLLM STDOUT: INFO 07-12 07:47:41 [api_server.py:1044] args: Namespace(subparser='serve', model_tag='unsloth/Llama-3.2-3B-Instruct', config='', host=None, port=8000, uvicorn_log_level='info', disable_uvicorn_access_log=False, allow_credentials=False, allowed_origins=['*'], allowed_methods=['*'], allowed_headers=['*'], api_key=None, lora_modules=None, prompt_adapters=None, chat_template=None, chat_template_content_format='auto', response_role='assistant', ssl_keyfile=None, ssl_certfile=None, ssl_ca_certs=None, enable_ssl_refresh=False, ssl_cert_reqs=0, root_path=None, middleware=[], return_tokens_as_token_ids=False, disable_frontend_multiprocessing=False, enable_request_id_headers=False, enable_auto_tool_choice=False, tool_call_parser=None, tool_parser_plugin='', model='unsloth/Llama-3.2-3B-Instruct', task='auto', tokenizer=None, hf_config_path=None, skip_tokenizer_init=False, revision=None, code_revision=None, tokenizer_revision=None, tokenizer_mode='auto', trust_remote_code=False, allowed_local_media_path=None, load_format='auto', download_dir=None, model_loader_extra_config={}, use_tqdm_on_load=True, config_format=<ConfigFormat.AUTO: 'auto'>, dtype='auto', max_model_len=2048, guided_decoding_backend='auto', reasoning_parser=None, logits_processor_pattern=None, model_impl='auto', distributed_executor_backend=None, pipeline_parallel_size=1, tensor_parallel_size=1, data_parallel_size=1, enable_expert_parallel=False, max_parallel_loading_workers=None, ray_workers_use_nsight=False, disable_custom_all_reduce=False, block_size=None, gpu_memory_utilization=0.8938626454842437, swap_space=4.0, kv_cache_dtype='auto', num_gpu_blocks_override=None, enable_prefix_caching=True, prefix_caching_hash_algo='builtin', cpu_offload_gb=0, calculate_kv_scales=False, disable_sliding_window=False, use_v2_block_manager=True, seed=0, max_logprobs=0, disable_log_stats=True, quantization=None, rope_scaling=None, rope_theta=None, hf_token=None, hf_overrides=None, enforce_eager=False, max_seq_len_to_capture=8192, tokenizer_pool_size=0, tokenizer_pool_type='ray', tokenizer_pool_extra_config={}, limit_mm_per_prompt={}, mm_processor_kwargs=None, disable_mm_preprocessor_cache=False, enable_lora=None, enable_lora_bias=False, max_loras=1, max_lora_rank=16, lora_extra_vocab_size=256, lora_dtype='auto', long_lora_scaling_factors=None, max_cpu_loras=None, fully_sharded_loras=False, enable_prompt_adapter=None, max_prompt_adapters=1, max_prompt_adapter_token=0, device='auto', speculative_config=None, ignore_patterns=[], served_model_name=None, qlora_adapter_name_or_path=None, show_hidden_metrics_for_version=None, otlp_traces_endpoint=None, collect_detailed_traces=None, disable_async_output_proc=False, max_num_batched_tokens=2048, max_num_seqs=192, max_num_partial_prefills=1, max_long_partial_prefills=1, long_prefill_token_threshold=0, num_lookahead_slots=0, scheduler_delay_factor=0.0, preemption_mode=None, num_scheduler_steps=1, multi_step_stream_outputs=True, scheduling_policy='fcfs', enable_chunked_prefill=None, disable_chunked_mm_input=False, scheduler_cls='vllm.core.scheduler.Scheduler', override_neuron_config=None, override_pooler_config=None, compilation_config={"level":3,"splitting_ops":[]}, kv_transfer_config=None, worker_cls='auto', worker_extension_cls='', generation_config='auto', override_generation_config=None, enable_sleep_mode=False, additional_config=None, enable_reasoning=False, disable_cascade_attn=False, disable_log_requests=False, max_log_len=None, disable_fastapi_docs=False, enable_prompt_tokens_details=False, enable_server_load_tracking=False, dispatch_function=<function ServeSubcommand.cmd at 0x785110c3bd80>)
vLLM STDOUT: INFO 07-12 07:47:55 [config.py:717] This model supports multiple tasks: {'embed', 'generate', 'score', 'classify', 'reward'}. Defaulting to 'generate'.
vLLM STDOUT: WARNING 07-12 07:47:55 [arg_utils.py:1658] Compute Capability < 8.0 is not supported by the V1 Engine. Falling back to V0.
vLLM STDOUT: INFO 07-12 07:47:55 [api_server.py:246] Started engine process with PID 2511
vLLM STDOUT: INFO 07-12 07:48:03 [__init__.py:239] Automatically detected platform cuda.
vLLM STDOUT: INFO 07-12 07:48:06 [llm_engine.py:240] Initializing a V0 LLM engine (v0.8.5.post1) with config: model='unsloth/Llama-3.2-3B-Instruct', speculative_config=None, tokenizer='unsloth/Llama-3.2-3B-Instruct', skip_tokenizer_init=False, tokenizer_mode=auto, revision=None, override_neuron_config=None, tokenizer_revision=None, trust_remote_code=False, dtype=torch.bfloat16, max_seq_len=2048, download_dir=None, load_format=LoadFormat.AUTO, tensor_parallel_size=1, pipeline_parallel_size=1, disable_custom_all_reduce=False, quantization=None, enforce_eager=False, kv_cache_dtype=auto, device_config=cuda, decoding_config=DecodingConfig(guided_decoding_backend='auto', reasoning_backend=None), observability_config=ObservabilityConfig(show_hidden_metrics=False, otlp_traces_endpoint=None, collect_model_forward_time=False, collect_model_execute_time=False), seed=0, served_model_name=unsloth/Llama-3.2-3B-Instruct, num_scheduler_steps=1, multi_step_stream_outputs=True, enable_prefix_caching=True, chunked_prefill_enabled=False, use_async_output_proc=True, disable_mm_preprocessor_cache=False, mm_processor_kwargs=None, pooler_config=None, compilation_config={"level":3,"splitting_ops":[],"compile_sizes":[],"cudagraph_capture_sizes":[192,184,176,168,160,152,144,136,128,120,112,104,96,88,80,72,64,56,48,40,32,24,16,8,4,2,1],"max_capture_size":192}, use_cached_outputs=True,
vLLM STDOUT: INFO 07-12 07:48:08 [cuda.py:240] Cannot use FlashAttention-2 backend for Volta and Turing GPUs.
vLLM STDOUT: INFO 07-12 07:48:08 [cuda.py:289] Using XFormers backend.
vLLM STDOUT: ERROR 07-12 07:48:08 [engine.py:448] Bfloat16 is only supported on GPUs with compute capability of at least 8.0. Your Tesla T4 GPU has compute capability 7.5. You can use float16 instead by explicitly setting the `dtype` flag in CLI, for example: --dtype=half.
vLLM STDOUT: ERROR 07-12 07:48:08 [engine.py:448] Traceback (most recent call last):
vLLM STDOUT: ERROR 07-12 07:48:08 [engine.py:448] File "/usr/local/lib/python3.11/dist-packages/vllm/engine/multiprocessing/engine.py", line 436, in run_mp_engine
vLLM STDOUT: ERROR 07-12 07:48:08 [engine.py:448] engine = MQLLMEngine.from_vllm_config(
vLLM STDOUT: ERROR 07-12 07:48:08 [engine.py:448] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
vLLM STDOUT: ERROR 07-12 07:48:08 [engine.py:448] File "/usr/local/lib/python3.11/dist-packages/vllm/engine/multiprocessing/engine.py", line 128, in from_vllm_config
vLLM STDOUT: ERROR 07-12 07:48:08 [engine.py:448] return cls(
vLLM STDOUT: ERROR 07-12 07:48:08 [engine.py:448] ^^^^
vLLM STDOUT: ERROR 07-12 07:48:08 [engine.py:448] File "/usr/local/lib/python3.11/dist-packages/vllm/engine/multiprocessing/engine.py", line 82, in __init__
vLLM STDOUT: ERROR 07-12 07:48:08 [engine.py:448] self.engine = LLMEngine(*args, **kwargs)
vLLM STDOUT: ERROR 07-12 07:48:08 [engine.py:448] ^^^^^^^^^^^^^^^^^^^^^^^^^^
vLLM STDOUT: ERROR 07-12 07:48:08 [engine.py:448] File "/usr/local/lib/python3.11/dist-packages/vllm/engine/llm_engine.py", line 275, in __init__
vLLM STDOUT: ERROR 07-12 07:48:08 [engine.py:448] self.model_executor = executor_class(vllm_config=vllm_config)
vLLM STDOUT: ERROR 07-12 07:48:08 [engine.py:448] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
vLLM STDOUT: ERROR 07-12 07:48:08 [engine.py:448] File "/usr/local/lib/python3.11/dist-packages/vllm/executor/executor_base.py", line 52, in __init__
vLLM STDOUT: ERROR 07-12 07:48:08 [engine.py:448] self._init_executor()
vLLM STDOUT: ERROR 07-12 07:48:08 [engine.py:448] File "/usr/local/lib/python3.11/dist-packages/vllm/executor/uniproc_executor.py", line 46, in _init_executor
vLLM STDOUT: ERROR 07-12 07:48:08 [engine.py:448] self.collective_rpc("init_device")
vLLM STDOUT: ERROR 07-12 07:48:08 [engine.py:448] File "/usr/local/lib/python3.11/dist-packages/vllm/executor/uniproc_executor.py", line 56, in collective_rpc
vLLM STDOUT: ERROR 07-12 07:48:08 [engine.py:448] answer = run_method(self.driver_worker, method, args, kwargs)
vLLM STDOUT: ERROR 07-12 07:48:08 [engine.py:448] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
vLLM STDOUT: ERROR 07-12 07:48:08 [engine.py:448] File "/usr/local/lib/python3.11/dist-packages/vllm/utils.py", line 2456, in run_method
vLLM STDOUT: ERROR 07-12 07:48:08 [engine.py:448] return func(*args, **kwargs)
vLLM STDOUT: ERROR 07-12 07:48:08 [engine.py:448] ^^^^^^^^^^^^^^^^^^^^^
vLLM STDOUT: ERROR 07-12 07:48:08 [engine.py:448] File "/usr/local/lib/python3.11/dist-packages/vllm/worker/worker_base.py", line 604, in init_device
vLLM STDOUT: ERROR 07-12 07:48:08 [engine.py:448] self.worker.init_device() # type: ignore
vLLM STDOUT: ERROR 07-12 07:48:08 [engine.py:448] ^^^^^^^^^^^^^^^^^^^^^^^^^
vLLM STDOUT: ERROR 07-12 07:48:08 [engine.py:448] File "/usr/local/lib/python3.11/dist-packages/vllm/worker/worker.py", line 177, in init_device
vLLM STDOUT: ERROR 07-12 07:48:08 [engine.py:448] _check_if_gpu_supports_dtype(self.model_config.dtype)
vLLM STDOUT: ERROR 07-12 07:48:08 [engine.py:448] File "/usr/local/lib/python3.11/dist-packages/vllm/worker/worker.py", line 546, in _check_if_gpu_supports_dtype
vLLM STDOUT: ERROR 07-12 07:48:08 [engine.py:448] raise ValueError(
vLLM STDOUT: ERROR 07-12 07:48:08 [engine.py:448] ValueError: Bfloat16 is only supported on GPUs with compute capability of at least 8.0. Your Tesla T4 GPU has compute capability 7.5. You can use float16 instead by explicitly setting the `dtype` flag in CLI, for example: --dtype=half.
Stdout stream ended before readiness message detected.