This repository has been archived by the owner on Oct 9, 2024. It is now read-only.
-
Notifications
You must be signed in to change notification settings - Fork 114
/
Makefile
144 lines (119 loc) · 4.74 KB
/
Makefile
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
gen-proto:
mkdir -p inference_server/model_handler/grpc_utils/pb
python -m grpc_tools.protoc -Iinference_server/model_handler/grpc_utils/proto --python_out=inference_server/model_handler/grpc_utils/pb --grpc_python_out=inference_server/model_handler/grpc_utils/pb inference_server/model_handler/grpc_utils/proto/generation.proto
find inference_server/model_handler/grpc_utils/pb/ -type f -name "*.py" -print0 -exec sed -i -e 's/^\(import.*pb2\)/from . \1/g' {} \;
touch inference_server/model_handler/grpc_utils/__init__.py
touch inference_server/model_handler/grpc_utils/pb/__init__.py
rm -rf inference_server/model_handler/grpc_utils/pb/*.py-e
ui:
python -m ui --ui_host 127.0.0.1 --ui_port 5001 --generation_backend_host 127.0.0.1 --generation_backend_port 5000 &
# ------------------------- DS inference -------------------------
bloom-176b:
make ui
TOKENIZERS_PARALLELISM=false \
MODEL_NAME=bigscience/bloom \
MODEL_CLASS=AutoModelForCausalLM \
DEPLOYMENT_FRAMEWORK=ds_inference \
DTYPE=fp16 \
MAX_INPUT_LENGTH=2048 \
MAX_BATCH_SIZE=4 \
CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7 \
gunicorn -t 0 -w 1 -b 127.0.0.1:5000 inference_server.server:app --access-logfile - --access-logformat '%(h)s %(t)s "%(r)s" %(s)s %(b)s'
# loads faster than the above one
microsoft-bloom-176b:
make ui
TOKENIZERS_PARALLELISM=false \
MODEL_NAME=microsoft/bloom-deepspeed-inference-fp16 \
MODEL_CLASS=AutoModelForCausalLM \
DEPLOYMENT_FRAMEWORK=ds_inference \
DTYPE=fp16 \
MAX_INPUT_LENGTH=2048 \
MAX_BATCH_SIZE=4 \
CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7 \
gunicorn -t 0 -w 1 -b 127.0.0.1:5000 inference_server.server:app --access-logfile - --access-logformat '%(h)s %(t)s "%(r)s" %(s)s %(b)s'
bloomz-176b:
make ui
TOKENIZERS_PARALLELISM=false \
MODEL_NAME=bigscience/bloomz \
MODEL_CLASS=AutoModelForCausalLM \
DEPLOYMENT_FRAMEWORK=ds_inference \
DTYPE=fp16 \
MAX_INPUT_LENGTH=2048 \
MAX_BATCH_SIZE=4 \
CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7 \
gunicorn -t 0 -w 1 -b 127.0.0.1:5000 inference_server.server:app --access-logfile - --access-logformat '%(h)s %(t)s "%(r)s" %(s)s %(b)s'
bloom-176b-int8:
make ui
TOKENIZERS_PARALLELISM=false \
MODEL_NAME=microsoft/bloom-deepspeed-inference-int8 \
MODEL_CLASS=AutoModelForCausalLM \
DEPLOYMENT_FRAMEWORK=ds_inference \
DTYPE=int8 \
MAX_INPUT_LENGTH=2048 \
MAX_BATCH_SIZE=4 \
CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7 \
gunicorn -t 0 -w 1 -b 127.0.0.1:5000 inference_server.server:app --access-logfile - --access-logformat '%(h)s %(t)s "%(r)s" %(s)s %(b)s'
# ------------------------- HF accelerate -------------------------
bloom-560m:
make ui
TOKENIZERS_PARALLELISM=false \
MODEL_NAME=bigscience/bloom-560m \
MODEL_CLASS=AutoModelForCausalLM \
DEPLOYMENT_FRAMEWORK=hf_accelerate \
DTYPE=bf16 \
MAX_INPUT_LENGTH=2048 \
MAX_BATCH_SIZE=32 \
CUDA_VISIBLE_DEVICES=0 \
gunicorn -t 0 -w 1 -b 127.0.0.1:5000 inference_server.server:app --access-logfile - --access-logformat '%(h)s %(t)s "%(r)s" %(s)s %(b)s'
flan-t5-xxl:
make ui
TOKENIZERS_PARALLELISM=false \
MODEL_NAME=google/flan-t5-xxl \
MODEL_CLASS=AutoModelForSeq2SeqLM \
DEPLOYMENT_FRAMEWORK=hf_accelerate \
DTYPE=bf16 \
MAX_INPUT_LENGTH=2048 \
MAX_BATCH_SIZE=4 \
CUDA_VISIBLE_DEVICES=0 \
gunicorn -t 0 -w 1 -b 127.0.0.1:5000 inference_server.server:app --access-logfile - --access-logformat '%(h)s %(t)s "%(r)s" %(s)s %(b)s'
ul2:
make ui
TOKENIZERS_PARALLELISM=false \
MODEL_NAME=google/ul2 \
MODEL_CLASS=AutoModelForSeq2SeqLM \
DEPLOYMENT_FRAMEWORK=hf_accelerate \
DTYPE=bf16 \
MAX_INPUT_LENGTH=2048 \
MAX_BATCH_SIZE=4 \
CUDA_VISIBLE_DEVICES=0 \
gunicorn -t 0 -w 1 -b 127.0.0.1:5000 inference_server.server:app --access-logfile - --access-logformat '%(h)s %(t)s "%(r)s" %(s)s %(b)s'
codegen-mono:
make ui
TOKENIZERS_PARALLELISM=false \
MODEL_NAME=Salesforce/codegen-16B-mono \
MODEL_CLASS=AutoModelForCausalLM \
DEPLOYMENT_FRAMEWORK=hf_accelerate \
DTYPE=bf16 \
MAX_INPUT_LENGTH=2048 \
MAX_BATCH_SIZE=4 \
CUDA_VISIBLE_DEVICES=0 \
gunicorn -t 0 -w 1 -b 127.0.0.1:5000 inference_server.server:app --access-logfile - --access-logformat '%(h)s %(t)s "%(r)s" %(s)s %(b)s'
# ------------------------- HF CPU -------------------------
bloom-560m-cpu:
make ui
MODEL_NAME=bigscience/bloom-560m \
MODEL_CLASS=AutoModelForCausalLM \
DEPLOYMENT_FRAMEWORK=hf_cpu \
DTYPE=fp32 \
MAX_INPUT_LENGTH=2048 \
MAX_BATCH_SIZE=32 \
gunicorn -t 0 -w 1 -b 127.0.0.1:5000 inference_server.server:app --access-logfile - --access-logformat '%(h)s %(t)s "%(r)s" %(s)s %(b)s'
flan-t5-base-cpu:
make ui
MODEL_NAME=google/flan-t5-base \
MODEL_CLASS=AutoModelForSeq2SeqLM \
DEPLOYMENT_FRAMEWORK=hf_cpu \
DTYPE=bf16 \
MAX_INPUT_LENGTH=2048 \
MAX_BATCH_SIZE=32 \
gunicorn -t 0 -w 1 -b 127.0.0.1:5000 inference_server.server:app --access-logfile - --access-logformat '%(h)s %(t)s "%(r)s" %(s)s %(b)s'