46 KiB
46 KiB
In [1]:
from sglang.utils import (
execute_shell_command,
wait_for_server,
terminate_process,
print_highlight,
)
server_process = execute_shell_command(
command="python -m sglang.launch_server --model-path meta-llama/Meta-Llama-3.1-8B-Instruct --port 30000 --host 0.0.0.0"
)
wait_for_server("http://localhost:30000")2024-10-30 09:44:20.477109: E external/local_xla/xla/stream_executor/cuda/cuda_fft.cc:479] Unable to register cuFFT factory: Attempting to register factory for plugin cuFFT when one has already been registered
2024-10-30 09:44:20.489679: E external/local_xla/xla/stream_executor/cuda/cuda_dnn.cc:10575] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered
2024-10-30 09:44:20.489712: E external/local_xla/xla/stream_executor/cuda/cuda_blas.cc:1442] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered
2024-10-30 09:44:21.010067: W tensorflow/compiler/tf2tensorrt/utils/py_utils.cc:38] TF-TRT Warning: Could not find TensorRT
[2024-10-30 09:44:29] server_args=ServerArgs(model_path='meta-llama/Meta-Llama-3.1-8B-Instruct', tokenizer_path='meta-llama/Meta-Llama-3.1-8B-Instruct', tokenizer_mode='auto', skip_tokenizer_init=False, load_format='auto', trust_remote_code=False, dtype='auto', kv_cache_dtype='auto', quantization=None, context_length=None, device='cuda', served_model_name='meta-llama/Meta-Llama-3.1-8B-Instruct', chat_template=None, is_embedding=False, host='0.0.0.0', port=30000, mem_fraction_static=0.88, max_running_requests=None, max_total_tokens=None, chunked_prefill_size=8192, max_prefill_tokens=16384, schedule_policy='lpm', schedule_conservativeness=1.0, tp_size=1, stream_interval=1, random_seed=134920821, constrained_json_whitespace_pattern=None, log_level='info', log_level_http=None, log_requests=False, show_time_cost=False, api_key=None, file_storage_pth='SGLang_storage', enable_cache_report=False, watchdog_timeout=600, dp_size=1, load_balance_method='round_robin', dist_init_addr=None, nnodes=1, node_rank=0, json_model_override_args='{}', enable_double_sparsity=False, ds_channel_config_path=None, ds_heavy_channel_num=32, ds_heavy_token_num=256, ds_heavy_channel_type='qk', ds_sparse_decode_threshold=4096, lora_paths=None, max_loras_per_batch=8, attention_backend='flashinfer', sampling_backend='flashinfer', grammar_backend='outlines', disable_flashinfer=False, disable_flashinfer_sampling=False, disable_radix_cache=False, disable_regex_jump_forward=False, disable_cuda_graph=False, disable_cuda_graph_padding=False, disable_disk_cache=False, disable_custom_all_reduce=False, disable_mla=False, disable_penalizer=False, disable_nan_detection=False, enable_overlap_schedule=False, enable_mixed_chunk=False, enable_torch_compile=False, torch_compile_max_bs=32, cuda_graph_max_bs=160, torchao_config='', enable_p2p_check=False, triton_attention_reduce_in_fp32=False, num_continuous_decode_steps=1)
[2024-10-30 09:44:39 TP0] Init torch distributed begin.
[2024-10-30 09:44:41 TP0] Load weight begin. avail mem=76.83 GB
[2024-10-30 09:44:42 TP0] lm_eval is not installed, GPTQ may not be usable
INFO 10-30 09:44:42 weight_utils.py:243] Using model weights format ['*.safetensors']
Loading safetensors checkpoint shards: 0% Completed | 0/4 [00:00<?, ?it/s]
Loading safetensors checkpoint shards: 25% Completed | 1/4 [00:01<00:05, 1.77s/it]
Loading safetensors checkpoint shards: 50% Completed | 2/4 [00:03<00:03, 1.77s/it]
Loading safetensors checkpoint shards: 75% Completed | 3/4 [00:05<00:01, 1.77s/it]
Loading safetensors checkpoint shards: 100% Completed | 4/4 [00:05<00:00, 1.27s/it]
Loading safetensors checkpoint shards: 100% Completed | 4/4 [00:05<00:00, 1.45s/it]
[2024-10-30 09:44:48 TP0] Load weight end. type=LlamaForCausalLM, dtype=torch.bfloat16, avail mem=61.82 GB
[2024-10-30 09:44:48 TP0] Memory pool end. avail mem=8.19 GB
[2024-10-30 09:44:49 TP0] Capture cuda graph begin. This can take up to several minutes.
[2024-10-30 09:44:58 TP0] max_total_num_tokens=430915, max_prefill_tokens=16384, max_running_requests=2049, context_len=131072
[2024-10-30 09:44:58] INFO: Started server process [231459]
[2024-10-30 09:44:58] INFO: Waiting for application startup.
[2024-10-30 09:44:58] INFO: Application startup complete.
[2024-10-30 09:44:58] INFO: Uvicorn running on http://0.0.0.0:30000 (Press CTRL+C to quit)
[2024-10-30 09:44:59] INFO: 127.0.0.1:54650 - "GET /v1/models HTTP/1.1" 200 OK
[2024-10-30 09:44:59] INFO: 127.0.0.1:54666 - "GET /get_model_info HTTP/1.1" 200 OK
[2024-10-30 09:44:59 TP0] Prefill batch. #new-seq: 1, #new-token: 7, #cached-token: 0, cache hit rate: 0.00%, token usage: 0.00, #running-req: 0, #queue-req: 0
[2024-10-30 09:44:59] INFO: 127.0.0.1:54672 - "POST /generate HTTP/1.1" 200 OK
[2024-10-30 09:44:59] The server is fired up and ready to roll!
NOTE: Typically, the server runs in a separate terminal.
In this notebook, we run the server and notebook code together, so their outputs are combined.
To improve clarity, the server logs are displayed in the original black color, while the notebook outputs are highlighted in blue.
In [2]:
import openai
client = openai.Client(base_url="http://127.0.0.1:30000/v1", api_key="None")
response = client.chat.completions.create(
model="meta-llama/Meta-Llama-3.1-8B-Instruct",
messages=[
{"role": "system", "content": "You are a helpful AI assistant"},
{"role": "user", "content": "List 3 countries and their capitals."},
],
temperature=0,
max_tokens=64,
)
print_highlight(f"Response: {response}")[2024-10-30 09:45:52 TP0] Prefill batch. #new-seq: 1, #new-token: 48, #cached-token: 1, cache hit rate: 1.79%, token usage: 0.00, #running-req: 0, #queue-req: 0 [2024-10-30 09:45:53 TP0] Decode batch. #running-req: 1, #token: 82, token usage: 0.00, gen throughput (token/s): 0.73, #queue-req: 0 [2024-10-30 09:45:53] INFO: 127.0.0.1:55594 - "POST /v1/chat/completions HTTP/1.1" 200 OK
Response: ChatCompletion(id='876500c402ae452ea17e4dde415c108a', choices=[Choice(finish_reason='stop', index=0, logprobs=None, message=ChatCompletionMessage(content='Here are 3 countries and their capitals:\n\n1. **Country:** Japan\n**Capital:** Tokyo\n\n2. **Country:** Australia\n**Capital:** Canberra\n\n3. **Country:** Brazil\n**Capital:** Brasília', refusal=None, role='assistant', audio=None, function_call=None, tool_calls=None), matched_stop=128009)], created=1730281553, model='meta-llama/Meta-Llama-3.1-8B-Instruct', object='chat.completion', service_tier=None, system_fingerprint=None, usage=CompletionUsage(completion_tokens=46, prompt_tokens=49, total_tokens=95, completion_tokens_details=None, prompt_tokens_details=None))
In [3]:
response = client.chat.completions.create(
model="meta-llama/Meta-Llama-3.1-8B-Instruct",
messages=[
{
"role": "system",
"content": "You are a knowledgeable historian who provides concise responses.",
},
{"role": "user", "content": "Tell me about ancient Rome"},
{
"role": "assistant",
"content": "Ancient Rome was a civilization centered in Italy.",
},
{"role": "user", "content": "What were their major achievements?"},
],
temperature=0.3, # Lower temperature for more focused responses
max_tokens=128, # Reasonable length for a concise response
top_p=0.95, # Slightly higher for better fluency
presence_penalty=0.2, # Mild penalty to avoid repetition
frequency_penalty=0.2, # Mild penalty for more natural language
n=1, # Single response is usually more stable
seed=42, # Keep for reproducibility
)
print_highlight(response.choices[0].message.content)[2024-10-30 09:45:57 TP0] Prefill batch. #new-seq: 1, #new-token: 48, #cached-token: 28, cache hit rate: 21.97%, token usage: 0.00, #running-req: 0, #queue-req: 0 [2024-10-30 09:45:57 TP0] Decode batch. #running-req: 1, #token: 104, token usage: 0.00, gen throughput (token/s): 8.70, #queue-req: 0 [2024-10-30 09:45:58 TP0] Decode batch. #running-req: 1, #token: 144, token usage: 0.00, gen throughput (token/s): 132.75, #queue-req: 0 [2024-10-30 09:45:58 TP0] Decode batch. #running-req: 1, #token: 184, token usage: 0.00, gen throughput (token/s): 132.30, #queue-req: 0 [2024-10-30 09:45:58] INFO: 127.0.0.1:55594 - "POST /v1/chat/completions HTTP/1.1" 200 OK
Ancient Rome's major achievements include:
1. **Engineering and Architecture**: Developed concrete, aqueducts, roads, bridges, and monumental buildings like the Colosseum and Pantheon.
2. **Law and Governance**: Established the Twelve Tables, a foundation for modern law, and a system of governance that included the Senate and Assemblies.
3. **Military Conquests**: Expanded the empire through numerous wars, creating a vast territory that stretched from Britain to Egypt.
4. **Language and Literature**: Developed Latin, which became the language of law, government, and literature, influencing modern languages like French, Spanish, and Italian.
1. **Engineering and Architecture**: Developed concrete, aqueducts, roads, bridges, and monumental buildings like the Colosseum and Pantheon.
2. **Law and Governance**: Established the Twelve Tables, a foundation for modern law, and a system of governance that included the Senate and Assemblies.
3. **Military Conquests**: Expanded the empire through numerous wars, creating a vast territory that stretched from Britain to Egypt.
4. **Language and Literature**: Developed Latin, which became the language of law, government, and literature, influencing modern languages like French, Spanish, and Italian.
In [4]:
stream = client.chat.completions.create(
model="meta-llama/Meta-Llama-3.1-8B-Instruct",
messages=[{"role": "user", "content": "Say this is a test"}],
stream=True,
)
for chunk in stream:
if chunk.choices[0].delta.content is not None:
print(chunk.choices[0].delta.content, end="")[2024-10-30 09:46:06] INFO: 127.0.0.1:45834 - "POST /v1/chat/completions HTTP/1.1" 200 OK [2024-10-30 09:46:06 TP0] Prefill batch. #new-seq: 1, #new-token: 15, #cached-token: 25, cache hit rate: 31.40%, token usage: 0.00, #running-req: 0, #queue-req: 0 It looks like you're getting started with our conversation. I'm happy to chat with you and see how[2024-10-30 09:46:06 TP0] Decode batch. #running-req: 1, #token: 61, token usage: 0.00, gen throughput (token/s): 4.78, #queue-req: 0 things go. What would you like to talk about?
In [5]:
response = client.completions.create(
model="meta-llama/Meta-Llama-3.1-8B-Instruct",
prompt="List 3 countries and their capitals.",
temperature=0,
max_tokens=64,
n=1,
stop=None,
)
print_highlight(f"Response: {response}")[2024-10-30 09:46:11 TP0] Prefill batch. #new-seq: 1, #new-token: 8, #cached-token: 1, cache hit rate: 30.39%, token usage: 0.00, #running-req: 0, #queue-req: 0 [2024-10-30 09:46:12 TP0] Decode batch. #running-req: 1, #token: 38, token usage: 0.00, gen throughput (token/s): 7.66, #queue-req: 0 [2024-10-30 09:46:12] INFO: 127.0.0.1:45834 - "POST /v1/completions HTTP/1.1" 200 OK
Response: Completion(id='1c988750627649f8872965d00cc008d9', choices=[CompletionChoice(finish_reason='length', index=0, logprobs=None, text=' 1. 2. 3.\n1. United States - Washington D.C. 2. Japan - Tokyo 3. Australia - Canberra\nList 3 countries and their capitals. 1. 2. 3.\n1. China - Beijing 2. Brazil - Bras', matched_stop=None)], created=1730281572, model='meta-llama/Meta-Llama-3.1-8B-Instruct', object='text_completion', system_fingerprint=None, usage=CompletionUsage(completion_tokens=64, prompt_tokens=9, total_tokens=73, completion_tokens_details=None, prompt_tokens_details=None))
In [6]:
response = client.completions.create(
model="meta-llama/Meta-Llama-3.1-8B-Instruct",
prompt="Write a short story about a space explorer.",
temperature=0.7, # Moderate temperature for creative writing
max_tokens=150, # Longer response for a story
top_p=0.9, # Balanced diversity in word choice
stop=["\n\n", "THE END"], # Multiple stop sequences
presence_penalty=0.3, # Encourage novel elements
frequency_penalty=0.3, # Reduce repetitive phrases
n=1, # Generate one completion
seed=123, # For reproducible results
)
print_highlight(f"Response: {response}")[2024-10-30 09:46:15 TP0] Prefill batch. #new-seq: 1, #new-token: 9, #cached-token: 1, cache hit rate: 29.32%, token usage: 0.00, #running-req: 0, #queue-req: 0 [2024-10-30 09:46:15 TP0] Decode batch. #running-req: 1, #token: 16, token usage: 0.00, gen throughput (token/s): 12.28, #queue-req: 0 [2024-10-30 09:46:15 TP0] Decode batch. #running-req: 1, #token: 56, token usage: 0.00, gen throughput (token/s): 135.70, #queue-req: 0 [2024-10-30 09:46:15 TP0] Decode batch. #running-req: 1, #token: 96, token usage: 0.00, gen throughput (token/s): 134.45, #queue-req: 0 [2024-10-30 09:46:16 TP0] Decode batch. #running-req: 1, #token: 136, token usage: 0.00, gen throughput (token/s): 133.34, #queue-req: 0 [2024-10-30 09:46:16] INFO: 127.0.0.1:45834 - "POST /v1/completions HTTP/1.1" 200 OK
Response: Completion(id='784041b9af634537a7960a0ba6152ba2', choices=[CompletionChoice(finish_reason='length', index=0, logprobs=None, text="\xa0\nOnce upon a time, in a distant corner of the universe, there was a brave space explorer named Captain Orion. She had spent her entire life studying the stars and dreaming of the day she could explore them for herself. Finally, after years of training and preparation, she set off on her maiden voyage to explore the cosmos.\nCaptain Orion's ship, the Aurora, was equipped with state-of-the-art technology and a crew of skilled astronauts who were eager to venture into the unknown. As they soared through the galaxy, they encountered breathtaking landscapes and incredible creatures that defied explanation.\nOn their first stop, they landed on a planet called Zorvath, a world of swirling purple clouds and towering crystal spires. Captain Orion and her crew mar", matched_stop=None)], created=1730281576, model='meta-llama/Meta-Llama-3.1-8B-Instruct', object='text_completion', system_fingerprint=None, usage=CompletionUsage(completion_tokens=150, prompt_tokens=10, total_tokens=160, completion_tokens_details=None, prompt_tokens_details=None))
In [6]:
import json
import time
from openai import OpenAI
client = OpenAI(base_url="http://127.0.0.1:30000/v1", api_key="None")
requests = [
{
"custom_id": "request-1",
"method": "POST",
"url": "/chat/completions",
"body": {
"model": "meta-llama/Meta-Llama-3.1-8B-Instruct",
"messages": [
{"role": "user", "content": "Tell me a joke about programming"}
],
"max_tokens": 50,
},
},
{
"custom_id": "request-2",
"method": "POST",
"url": "/chat/completions",
"body": {
"model": "meta-llama/Meta-Llama-3.1-8B-Instruct",
"messages": [{"role": "user", "content": "What is Python?"}],
"max_tokens": 50,
},
},
]
input_file_path = "batch_requests.jsonl"
with open(input_file_path, "w") as f:
for req in requests:
f.write(json.dumps(req) + "\n")
with open(input_file_path, "rb") as f:
file_response = client.files.create(file=f, purpose="batch")
batch_response = client.batches.create(
input_file_id=file_response.id,
endpoint="/v1/chat/completions",
completion_window="24h",
)
print_highlight(f"Batch job created with ID: {batch_response.id}")[2024-10-28 02:02:55] INFO: 127.0.0.1:43330 - "POST /v1/files HTTP/1.1" 200 OK [2024-10-28 02:02:55] INFO: 127.0.0.1:43330 - "POST /v1/batches HTTP/1.1" 200 OK [2024-10-28 02:02:55 TP0] Prefill batch. #new-seq: 2, #new-token: 30, #cached-token: 50, cache hit rate: 35.06%, token usage: 0.00, #running-req: 0, #queue-req: 0
Batch job created with ID: batch_56fefd2e-0187-4c14-aa2d-110917723dde
In [7]:
while batch_response.status not in ["completed", "failed", "cancelled"]:
time.sleep(3)
print(f"Batch job status: {batch_response.status}...trying again in 3 seconds...")
batch_response = client.batches.retrieve(batch_response.id)
if batch_response.status == "completed":
print("Batch job completed successfully!")
print(f"Request counts: {batch_response.request_counts}")
result_file_id = batch_response.output_file_id
file_response = client.files.content(result_file_id)
result_content = file_response.read().decode("utf-8")
results = [
json.loads(line) for line in result_content.split("\n") if line.strip() != ""
]
for result in results:
print_highlight(f"Request {result['custom_id']}:")
print_highlight(f"Response: {result['response']}")
print_highlight("Cleaning up files...")
# Only delete the result file ID since file_response is just content
client.files.delete(result_file_id)
else:
print_highlight(f"Batch job failed with status: {batch_response.status}")
if hasattr(batch_response, "errors"):
print_highlight(f"Errors: {batch_response.errors}")[2024-10-28 02:02:56 TP0] Decode batch. #running-req: 2, #token: 82, token usage: 0.00, gen throughput (token/s): 55.10, #queue-req: 0 Batch job status: validating...trying again in 3 seconds... [2024-10-28 02:02:58] INFO: 127.0.0.1:43330 - "GET /v1/batches/batch_56fefd2e-0187-4c14-aa2d-110917723dde HTTP/1.1" 200 OK Batch job completed successfully! Request counts: BatchRequestCounts(completed=2, failed=0, total=2) [2024-10-28 02:02:58] INFO: 127.0.0.1:43330 - "GET /v1/files/backend_result_file-520da6c8-0cce-4d4c-a943-a86101f5f5b4/content HTTP/1.1" 200 OK
Request request-1:
Response: {'status_code': 200, 'request_id': 'request-1', 'body': {'id': 'request-1', 'object': 'chat.completion', 'created': 1730106176, 'model': 'meta-llama/Meta-Llama-3.1-8B-Instruct', 'choices': {'index': 0, 'message': {'role': 'assistant', 'content': 'A programmer walks into a library and asks the librarian, "Do you have any books on Pavlov\'s dogs and Schrödinger\'s cat?"\n\nThe librarian replies, "It rings a bell, but I\'m not sure if it\'s here'}, 'logprobs': None, 'finish_reason': 'length', 'matched_stop': None}, 'usage': {'prompt_tokens': 41, 'completion_tokens': 50, 'total_tokens': 91}, 'system_fingerprint': None}}
Request request-2:
Response: {'status_code': 200, 'request_id': 'request-2', 'body': {'id': 'request-2', 'object': 'chat.completion', 'created': 1730106176, 'model': 'meta-llama/Meta-Llama-3.1-8B-Instruct', 'choices': {'index': 0, 'message': {'role': 'assistant', 'content': '**What is Python?**\n\nPython is a high-level, interpreted programming language that is widely used for various purposes, including:\n\n1. **Web Development**: Building web applications and web services using frameworks like Django and Flask.\n2. **Data Analysis and'}, 'logprobs': None, 'finish_reason': 'length', 'matched_stop': None}, 'usage': {'prompt_tokens': 39, 'completion_tokens': 50, 'total_tokens': 89}, 'system_fingerprint': None}}
Cleaning up files...
[2024-10-28 02:02:58] INFO: 127.0.0.1:43330 - "DELETE /v1/files/backend_result_file-520da6c8-0cce-4d4c-a943-a86101f5f5b4 HTTP/1.1" 200 OK
In [8]:
import json
import time
from openai import OpenAI
client = OpenAI(base_url="http://127.0.0.1:30000/v1", api_key="None")
requests = []
for i in range(100):
requests.append(
{
"custom_id": f"request-{i}",
"method": "POST",
"url": "/chat/completions",
"body": {
"model": "meta-llama/Meta-Llama-3.1-8B-Instruct",
"messages": [
{
"role": "system",
"content": f"{i}: You are a helpful AI assistant",
},
{
"role": "user",
"content": "Write a detailed story about topic. Make it very long.",
},
],
"max_tokens": 500,
},
}
)
input_file_path = "batch_requests.jsonl"
with open(input_file_path, "w") as f:
for req in requests:
f.write(json.dumps(req) + "\n")
with open(input_file_path, "rb") as f:
uploaded_file = client.files.create(file=f, purpose="batch")
batch_job = client.batches.create(
input_file_id=uploaded_file.id,
endpoint="/v1/chat/completions",
completion_window="24h",
)
print_highlight(f"Created batch job with ID: {batch_job.id}")
print_highlight(f"Initial status: {batch_job.status}")
time.sleep(10)
max_checks = 5
for i in range(max_checks):
batch_details = client.batches.retrieve(batch_id=batch_job.id)
print_highlight(
f"Batch job details (check {i+1} / {max_checks}) // ID: {batch_details.id} // Status: {batch_details.status} // Created at: {batch_details.created_at} // Input file ID: {batch_details.input_file_id} // Output file ID: {batch_details.output_file_id}"
)
print_highlight(
f"<strong>Request counts: Total: {batch_details.request_counts.total} // Completed: {batch_details.request_counts.completed} // Failed: {batch_details.request_counts.failed}</strong>"
)
time.sleep(3)[2024-10-28 02:02:58] INFO: 127.0.0.1:43336 - "POST /v1/files HTTP/1.1" 200 OK [2024-10-28 02:02:58] INFO: 127.0.0.1:43336 - "POST /v1/batches HTTP/1.1" 200 OK
Created batch job with ID: batch_67da0e16-e7b2-4a75-9f7a-58c033e739e5
Initial status: validating
[2024-10-28 02:02:58 TP0] Prefill batch. #new-seq: 17, #new-token: 510, #cached-token: 425, cache hit rate: 43.40%, token usage: 0.00, #running-req: 0, #queue-req: 0 [2024-10-28 02:02:58 TP0] Prefill batch. #new-seq: 83, #new-token: 2490, #cached-token: 2075, cache hit rate: 45.04%, token usage: 0.00, #running-req: 17, #queue-req: 0 [2024-10-28 02:02:59 TP0] Decode batch. #running-req: 100, #token: 3725, token usage: 0.02, gen throughput (token/s): 234.43, #queue-req: 0 [2024-10-28 02:03:00 TP0] Decode batch. #running-req: 100, #token: 7725, token usage: 0.04, gen throughput (token/s): 3545.41, #queue-req: 0 [2024-10-28 02:03:01 TP0] Decode batch. #running-req: 100, #token: 11725, token usage: 0.05, gen throughput (token/s): 3448.10, #queue-req: 0 [2024-10-28 02:03:02 TP0] Decode batch. #running-req: 100, #token: 15725, token usage: 0.07, gen throughput (token/s): 3362.62, #queue-req: 0 [2024-10-28 02:03:04 TP0] Decode batch. #running-req: 100, #token: 19725, token usage: 0.09, gen throughput (token/s): 3279.58, #queue-req: 0 [2024-10-28 02:03:05 TP0] Decode batch. #running-req: 100, #token: 23725, token usage: 0.11, gen throughput (token/s): 3200.86, #queue-req: 0 [2024-10-28 02:03:06 TP0] Decode batch. #running-req: 100, #token: 27725, token usage: 0.13, gen throughput (token/s): 3126.52, #queue-req: 0 [2024-10-28 02:03:07 TP0] Decode batch. #running-req: 100, #token: 31725, token usage: 0.15, gen throughput (token/s): 3053.16, #queue-req: 0 [2024-10-28 02:03:08] INFO: 127.0.0.1:41320 - "GET /v1/batches/batch_67da0e16-e7b2-4a75-9f7a-58c033e739e5 HTTP/1.1" 200 OK
Batch job details (check 1 / 5) // ID: batch_67da0e16-e7b2-4a75-9f7a-58c033e739e5 // Status: in_progress // Created at: 1730106178 // Input file ID: backend_input_file-92cf2cc1-afbd-428f-8c5c-85fabd86cb63 // Output file ID: None
Request counts: Total: 0 // Completed: 0 // Failed: 0
[2024-10-28 02:03:09 TP0] Decode batch. #running-req: 100, #token: 35725, token usage: 0.16, gen throughput (token/s): 2980.26, #queue-req: 0 [2024-10-28 02:03:10 TP0] Decode batch. #running-req: 100, #token: 39725, token usage: 0.18, gen throughput (token/s): 2919.09, #queue-req: 0 [2024-10-28 02:03:11] INFO: 127.0.0.1:41320 - "GET /v1/batches/batch_67da0e16-e7b2-4a75-9f7a-58c033e739e5 HTTP/1.1" 200 OK
Batch job details (check 2 / 5) // ID: batch_67da0e16-e7b2-4a75-9f7a-58c033e739e5 // Status: in_progress // Created at: 1730106178 // Input file ID: backend_input_file-92cf2cc1-afbd-428f-8c5c-85fabd86cb63 // Output file ID: None
Request counts: Total: 0 // Completed: 0 // Failed: 0
[2024-10-28 02:03:11 TP0] Decode batch. #running-req: 100, #token: 43725, token usage: 0.20, gen throughput (token/s): 2854.92, #queue-req: 0 [2024-10-28 02:03:13 TP0] Decode batch. #running-req: 100, #token: 47725, token usage: 0.22, gen throughput (token/s): 2794.62, #queue-req: 0 [2024-10-28 02:03:14] INFO: 127.0.0.1:41320 - "GET /v1/batches/batch_67da0e16-e7b2-4a75-9f7a-58c033e739e5 HTTP/1.1" 200 OK
Batch job details (check 3 / 5) // ID: batch_67da0e16-e7b2-4a75-9f7a-58c033e739e5 // Status: in_progress // Created at: 1730106178 // Input file ID: backend_input_file-92cf2cc1-afbd-428f-8c5c-85fabd86cb63 // Output file ID: None
Request counts: Total: 0 // Completed: 0 // Failed: 0
[2024-10-28 02:03:14 TP0] Decode batch. #running-req: 100, #token: 51725, token usage: 0.24, gen throughput (token/s): 2737.84, #queue-req: 0 [2024-10-28 02:03:17] INFO: 127.0.0.1:41320 - "GET /v1/batches/batch_67da0e16-e7b2-4a75-9f7a-58c033e739e5 HTTP/1.1" 200 OK
Batch job details (check 4 / 5) // ID: batch_67da0e16-e7b2-4a75-9f7a-58c033e739e5 // Status: completed // Created at: 1730106178 // Input file ID: backend_input_file-92cf2cc1-afbd-428f-8c5c-85fabd86cb63 // Output file ID: backend_result_file-c10ee9f5-eca8-4357-a922-934543b7f433
Request counts: Total: 100 // Completed: 100 // Failed: 0
[2024-10-28 02:03:20] INFO: 127.0.0.1:41320 - "GET /v1/batches/batch_67da0e16-e7b2-4a75-9f7a-58c033e739e5 HTTP/1.1" 200 OK
Batch job details (check 5 / 5) // ID: batch_67da0e16-e7b2-4a75-9f7a-58c033e739e5 // Status: completed // Created at: 1730106178 // Input file ID: backend_input_file-92cf2cc1-afbd-428f-8c5c-85fabd86cb63 // Output file ID: backend_result_file-c10ee9f5-eca8-4357-a922-934543b7f433
Request counts: Total: 100 // Completed: 100 // Failed: 0
In [9]:
import json
import time
from openai import OpenAI
client = OpenAI(base_url="http://127.0.0.1:30000/v1", api_key="None")
requests = []
for i in range(500):
requests.append(
{
"custom_id": f"request-{i}",
"method": "POST",
"url": "/chat/completions",
"body": {
"model": "meta-llama/Meta-Llama-3.1-8B-Instruct",
"messages": [
{
"role": "system",
"content": f"{i}: You are a helpful AI assistant",
},
{
"role": "user",
"content": "Write a detailed story about topic. Make it very long.",
},
],
"max_tokens": 500,
},
}
)
input_file_path = "batch_requests.jsonl"
with open(input_file_path, "w") as f:
for req in requests:
f.write(json.dumps(req) + "\n")
with open(input_file_path, "rb") as f:
uploaded_file = client.files.create(file=f, purpose="batch")
batch_job = client.batches.create(
input_file_id=uploaded_file.id,
endpoint="/v1/chat/completions",
completion_window="24h",
)
print_highlight(f"Created batch job with ID: {batch_job.id}")
print_highlight(f"Initial status: {batch_job.status}")
time.sleep(10)
try:
cancelled_job = client.batches.cancel(batch_id=batch_job.id)
print_highlight(f"Cancellation initiated. Status: {cancelled_job.status}")
assert cancelled_job.status == "cancelling"
# Monitor the cancellation process
while cancelled_job.status not in ["failed", "cancelled"]:
time.sleep(3)
cancelled_job = client.batches.retrieve(batch_job.id)
print_highlight(f"Current status: {cancelled_job.status}")
# Verify final status
assert cancelled_job.status == "cancelled"
print_highlight("Batch job successfully cancelled")
except Exception as e:
print_highlight(f"Error during cancellation: {e}")
raise e
finally:
try:
del_response = client.files.delete(uploaded_file.id)
if del_response.deleted:
print_highlight("Successfully cleaned up input file")
except Exception as e:
print_highlight(f"Error cleaning up: {e}")
raise e[2024-10-28 02:03:23] INFO: 127.0.0.1:47360 - "POST /v1/files HTTP/1.1" 200 OK [2024-10-28 02:03:23] INFO: 127.0.0.1:47360 - "POST /v1/batches HTTP/1.1" 200 OK
Created batch job with ID: batch_8a409f86-b8c7-4e29-9cc7-187d6d28df62
Initial status: validating
[2024-10-28 02:03:23 TP0] Prefill batch. #new-seq: 44, #new-token: 44, #cached-token: 2376, cache hit rate: 60.81%, token usage: 0.01, #running-req: 0, #queue-req: 0 [2024-10-28 02:03:23 TP0] Prefill batch. #new-seq: 328, #new-token: 8192, #cached-token: 9824, cache hit rate: 56.49%, token usage: 0.01, #running-req: 44, #queue-req: 128 [2024-10-28 02:03:24 TP0] Prefill batch. #new-seq: 129, #new-token: 3864, #cached-token: 3231, cache hit rate: 54.15%, token usage: 0.05, #running-req: 371, #queue-req: 1 [2024-10-28 02:03:27 TP0] Decode batch. #running-req: 500, #token: 29025, token usage: 0.13, gen throughput (token/s): 1162.55, #queue-req: 0 [2024-10-28 02:03:31 TP0] Decode batch. #running-req: 500, #token: 49025, token usage: 0.23, gen throughput (token/s): 5606.35, #queue-req: 0 [2024-10-28 02:03:33] INFO: 127.0.0.1:40110 - "POST /v1/batches/batch_8a409f86-b8c7-4e29-9cc7-187d6d28df62/cancel HTTP/1.1" 200 OK
Cancellation initiated. Status: cancelling
[2024-10-28 02:03:36] INFO: 127.0.0.1:40110 - "GET /v1/batches/batch_8a409f86-b8c7-4e29-9cc7-187d6d28df62 HTTP/1.1" 200 OK
Current status: cancelled
Batch job successfully cancelled
[2024-10-28 02:03:36] INFO: 127.0.0.1:40110 - "DELETE /v1/files/backend_input_file-2e9608b6-981b-48ec-8adb-e653ffc69106 HTTP/1.1" 200 OK
Successfully cleaned up input file
In [7]:
terminate_process(server_process)