Skip to content

GitLab

"vscode:/vscode.git/clone" did not exist on "9d5b4e4deaa3318df49419d325490730391efd75"
Unverified Commit 46004e83 authored by youkaichao's avatar youkaichao Committed by GitHub
Browse files

[misc] clean up and unify logging (#10999) 


Signed-off-by: default avataryoukaichao <youkaichao@gmail.com>
parent 43b05fa3
Show whitespace changes
Inline Side-by-side
Showing with 37 additions and 90 deletions
vllm/config.py
View file @ 46004e83
...@@ -2579,45 +2579,40 @@ class VllmConfig: ...@@ -2579,45 +2579,40 @@ class VllmConfig:
self.instance_id = random_uuid()[:5] self.instance_id = random_uuid()[:5]
def __str__(self): def __str__(self):
return ("model=%r, speculative_config=%r, tokenizer=%r, " return (
"skip_tokenizer_init=%s, tokenizer_mode=%s, revision=%s, " f"model={self.model_config.model!r},"
"override_neuron_config=%s, tokenizer_revision=%s, " f" speculative_config={self.speculative_config!r},"
"trust_remote_code=%s, dtype=%s, max_seq_len=%d, " f" tokenizer={self.model_config.tokenizer!r}, "
"download_dir=%r, load_format=%s, tensor_parallel_size=%d, " f"skip_tokenizer_init={self.model_config.skip_tokenizer_init},"
"pipeline_parallel_size=%d, " f" tokenizer_mode={self.model_config.tokenizer_mode}, "
"disable_custom_all_reduce=%s, quantization=%s, " f"revision={self.model_config.revision}, "
"enforce_eager=%s, kv_cache_dtype=%s, " f"override_neuron_config={self.model_config.override_neuron_config},"
"quantization_param_path=%s, device_config=%s, " f" tokenizer_revision={self.model_config.tokenizer_revision}, "
"decoding_config=%r, observability_config=%r, " f"trust_remote_code={self.model_config.trust_remote_code}, "
"seed=%d, served_model_name=%s, " f"dtype={self.model_config.dtype}, "
"num_scheduler_steps=%d, enable_prefix_caching=%s, " f"max_seq_len={self.model_config.max_model_len},"
"use_async_output_proc=%s, mm_processor_kwargs=%s") % \ f" download_dir={self.load_config.download_dir!r}, "
(self.model_config.model, self.speculative_config, f"load_format={self.load_config.load_format}, "
self.model_config.tokenizer, f"tensor_parallel_size={self.parallel_config.tensor_parallel_size},"
self.model_config.skip_tokenizer_init, f" pipeline_parallel_size={self.parallel_config.pipeline_parallel_size}, " # noqa
self.model_config.tokenizer_mode, f"disable_custom_all_reduce={self.parallel_config.disable_custom_all_reduce}, " # noqa
self.model_config.revision, f"quantization={self.model_config.quantization}, "
self.model_config.override_neuron_config, f"enforce_eager={self.model_config.enforce_eager}, "
self.model_config.tokenizer_revision, f"kv_cache_dtype={self.cache_config.cache_dtype}, "
self.model_config.trust_remote_code, f"quantization_param_path={self.model_config.quantization_param_path},"
self.model_config.dtype, f" device_config={self.device_config.device}, "
self.model_config.max_model_len, f"decoding_config={self.decoding_config!r}, "
self.load_config.download_dir, f"observability_config={self.observability_config!r}, "
self.load_config.load_format, f"seed={self.model_config.seed}, "
self.parallel_config.tensor_parallel_size, f"served_model_name={self.model_config.served_model_name}, "
self.parallel_config.pipeline_parallel_size, f"num_scheduler_steps={self.scheduler_config.num_scheduler_steps}, "
self.parallel_config.disable_custom_all_reduce, f"multi_step_stream_outputs={self.scheduler_config.multi_step_stream_outputs}, " # noqa
self.model_config.quantization, f"enable_prefix_caching={self.cache_config.enable_prefix_caching}, "
self.model_config.enforce_eager, f"chunked_prefill_enabled={self.scheduler_config.chunked_prefill_enabled}, " # noqa
self.cache_config.cache_dtype, f"use_async_output_proc={self.model_config.use_async_output_proc}, "
self.model_config.quantization_param_path, f"mm_processor_kwargs={self.model_config.mm_processor_kwargs}, "
self.device_config.device, self.decoding_config, f"pooler_config={self.model_config.pooler_config!r},"
self.observability_config, self.model_config.seed, f" compilation_config={self.compilation_config!r}")
self.model_config.served_model_name,
self.scheduler_config.num_scheduler_steps,
self.cache_config.enable_prefix_caching,
self.model_config.use_async_output_proc,
self.model_config.mm_processor_kwargs)
_current_vllm_config: Optional[VllmConfig] = None _current_vllm_config: Optional[VllmConfig] = None
......
vllm/engine/llm_engine.py
View file @ 46004e83
...@@ -247,60 +247,12 @@ class LLMEngine: ...@@ -247,60 +247,12 @@ class LLMEngine:
) )
logger.info( logger.info(
"Initializing an LLM engine (v%s) with config: " "Initializing an LLM engine (v%s) with config: %r,"
"model=%r, speculative_config=%r, tokenizer=%r, " "use_cached_outputs=%s, ",
"skip_tokenizer_init=%s, tokenizer_mode=%s, revision=%s, "
"override_neuron_config=%s, tokenizer_revision=%s, "
"trust_remote_code=%s, dtype=%s, max_seq_len=%d, "
"download_dir=%r, load_format=%s, tensor_parallel_size=%d, "
"pipeline_parallel_size=%d, "
"disable_custom_all_reduce=%s, quantization=%s, "
"enforce_eager=%s, kv_cache_dtype=%s, "
"quantization_param_path=%s, device_config=%s, "
"decoding_config=%r, observability_config=%r, "
"seed=%d, served_model_name=%s, "
"num_scheduler_steps=%d, chunked_prefill_enabled=%s "
"multi_step_stream_outputs=%s, enable_prefix_caching=%s, "
"use_async_output_proc=%s, use_cached_outputs=%s, "
"mm_processor_kwargs=%s, pooler_config=%r,"
"compilation_config=%r",
VLLM_VERSION, VLLM_VERSION,
self.model_config.model, vllm_config,
self.speculative_config,
self.model_config.tokenizer,
self.model_config.skip_tokenizer_init,
self.model_config.tokenizer_mode,
self.model_config.revision,
self.model_config.override_neuron_config,
self.model_config.tokenizer_revision,
self.model_config.trust_remote_code,
self.model_config.dtype,
self.model_config.max_model_len,
self.load_config.download_dir,
self.load_config.load_format,
self.parallel_config.tensor_parallel_size,
self.parallel_config.pipeline_parallel_size,
self.parallel_config.disable_custom_all_reduce,
self.model_config.quantization,
self.model_config.enforce_eager,
self.cache_config.cache_dtype,
self.model_config.quantization_param_path,
self.device_config.device,
self.decoding_config,
self.observability_config,
self.model_config.seed,
self.model_config.served_model_name,
self.scheduler_config.num_scheduler_steps,
self.scheduler_config.chunked_prefill_enabled,
self.scheduler_config.multi_step_stream_outputs,
self.cache_config.enable_prefix_caching,
self.model_config.use_async_output_proc,
use_cached_outputs, use_cached_outputs,
self.model_config.mm_processor_kwargs,
self.model_config.pooler_config,
vllm_config.compilation_config,
) )
# TODO(woosuk): Print more configs in debug mode.
self.log_stats = log_stats self.log_stats = log_stats
self.use_cached_outputs = use_cached_outputs self.use_cached_outputs = use_cached_outputs
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment