mirror/OpenLLM
1
0
Fork 0
mirror of https://github.com/bentoml/OpenLLM.git synced 2026-05-19 14:16:22 -04:00
Code Issues Packages Projects Releases Wiki Activity

refactor: alias

Browse Source
This commit is contained in:
bojiang
2024-07-02 22:20:02 +08:00
parent 5d914aa2bc
commit 016b7b016d
1 changed files with 22 additions and 64 deletions
Download Patch File Download Diff File Expand all files Collapse all files

86
recipe.yaml
View File

@@ -1,6 +1,4 @@
"phi3:3.8b-mini-instruct-4k-fp16":
alias:
- 3.8b
"phi3:3.8b-instruct-fp16":
project: vllm-chat
service_config:
name: phi3
@@ -15,11 +13,9 @@
dtype: half
chat_template: phi-3
extra_labels:
openllm_alias: 3.8b
openllm_alias: 3.8b,3.8b-mini,3.8b-mini-instruct-4k-fp16
openllm_hf_model_id: microsoft/Phi-3-mini-4k-instruct
"llama2:7b-chat-fp16":
alias:
- 7b
project: vllm-chat
service_config:
name: llama2
@@ -34,11 +30,9 @@
dtype: half
chat_template: llama-2-chat
extra_labels:
openllm_alias: 7b
openllm_alias: 7b,7b-chat
openllm_hf_model_id: meta-llama/Llama-2-7b-chat-hf
"llama2:13b-chat-fp16":
alias:
- 13b
project: vllm-chat
service_config:
name: llama2
@@ -53,11 +47,9 @@
dtype: half
chat_template: llama-2-chat
extra_labels:
openllm_alias: 13b
openllm_alias: 13b,13b-chat
openllm_hf_model_id: meta-llama/Llama-2-13b-chat-hf
"llama2:70b-chat-fp16":
alias:
- 70b
project: vllm-chat
service_config:
name: llama2
@@ -72,11 +64,9 @@
dtype: half
chat_template: llama-2-chat
extra_labels:
openllm_alias: 70b
openllm_alias: 70b,70b-chat
openllm_hf_model_id: meta-llama/Llama-2-70b-chat-hf
"llama2:7b-chat-awq-4bit":
alias:
- 7b-4bit
project: vllm-chat
service_config:
name: llama2
@@ -92,11 +82,9 @@
enforce_eager: true
chat_template: llama-2-chat
extra_labels:
openllm_alias: 7b-4bit
openllm_alias: 7b-4bit,7b-chat-4bit
openllm_hf_model_id: TheBloke/Llama-2-7B-Chat-AWQ
"mistral:7b-instruct-awq-4bit":
alias:
- 7b-4bit
project: vllm-chat
service_config:
name: mistral
@@ -113,11 +101,9 @@
dtype: half
chat_template: mistral-instruct
extra_labels:
openllm_alias: 7b-4bit
openllm_alias: 7b-4bit,7b-instruct-4bit
openllm_hf_model_id: TheBloke/Mistral-7B-Instruct-v0.1-AWQ
"mistral:7b-instruct-fp16":
alias:
- 7b
project: vllm-chat
service_config:
name: mistral
@@ -133,11 +119,9 @@
dtype: half
chat_template: mistral-instruct
extra_labels:
openllm_alias: 7b
openllm_alias: 7b,7b-instruct
openllm_hf_model_id: mistralai/Mistral-7B-Instruct-v0.1
"llama3:8b-instruct-awq-4bit":
alias:
- 8b-4bit
project: vllm-chat
service_config:
name: llama3
@@ -151,11 +135,9 @@
max_model_len: 2048
quantization: awq
extra_labels:
openllm_alias: 8b-4bit
openllm_alias: 8b-4bit,8b-instruct-4bit
openllm_hf_model_id: casperhansen/llama-3-8b-instruct-awq
"llama3:70b-instruct-awq-4bit":
alias:
- 70b-4bit
project: vllm-chat
service_config:
name: llama3
@@ -169,11 +151,9 @@
max_model_len: 2048
quantization: awq
extra_labels:
openllm_alias: 70b-4bit
openllm_alias: 70b-4bit,70b-instruct-4bit
openllm_hf_model_id: casperhansen/llama-3-70b-instruct-awq
"llama3:8b-instruct-fp16":
alias:
- 8b
project: vllm-chat
service_config:
name: llama3
@@ -187,11 +167,9 @@
max_model_len: 2048
dtype: half
extra_labels:
openllm_alias: 8b
openllm_alias: 8b,8b-instruct
openllm_hf_model_id: meta-llama/Meta-Llama-3-8B-Instruct
"llama3:70b-instruct-fp16":
alias:
- 70b
project: vllm-chat
service_config:
name: llama3
@@ -204,11 +182,9 @@
model: meta-llama/Meta-Llama-3-70B-Instruct
max_model_len: 2048
extra_labels:
openllm_alias: 70b
openllm_alias: 70b,70b-instruct
openllm_hf_model_id: meta-llama/Meta-Llama-3-70B-Instruct
"gemma:2b-instruct-fp16":
alias:
- 2b
project: vllm-chat
service_config:
name: gemma
@@ -222,11 +198,9 @@
max_model_len: 2048
dtype: half
extra_labels:
openllm_alias: 2b
openllm_alias: 2b,2b-instruct
openllm_hf_model_id: google/gemma-2b-it
"gemma:7b-instruct-fp16":
alias:
- 7b
project: vllm-chat
service_config:
name: gemma
@@ -240,11 +214,9 @@
max_model_len: 2048
dtype: half
extra_labels:
openllm_alias: 7b
openllm_alias: 7b,7b-instruct
openllm_hf_model_id: google/gemma-7b-it
"gemma:7b-instruct-awq-4bit":
alias:
- 7b-4bit
project: vllm-chat
service_config:
name: gemma
@@ -259,7 +231,7 @@
quantization: awq
chat_template: gemma-it
extra_labels:
openllm_alias: 7b-4bit
openllm_alias: 7b-4bit,7b-instruct-4bit
openllm_hf_model_id: casperhansen/gemma-7b-it-awq
"mixtral:8x7b-instruct-v0.1-fp16":
project: vllm-chat
@@ -296,8 +268,6 @@
openllm_alias: 8x7b-4bit
openllm_hf_model_id: casperhansen/mixtral-instruct-awq
"qwen2:0.5b-instruct-fp16":
alias:
- 0.5b
project: vllm-chat
service_config:
name: qwen2
@@ -311,11 +281,9 @@
max_model_len: 2048
dtype: half
extra_labels:
openllm_alias: 0.5b
openllm_alias: 0.5b,0.5b-instruct
openllm_hf_model_id: Qwen/Qwen2-0.5B-Instruct
"qwen2:1.5b-instruct-fp16":
alias:
- 1.5b
project: vllm-chat
service_config:
name: qwen2
@@ -329,11 +297,9 @@
max_model_len: 2048
dtype: half
extra_labels:
openllm_alias: 1.5b
openllm_alias: 1.5b,1.5b-instruct
openllm_hf_model_id: Qwen/Qwen2-1.5B-Instruct
"qwen2:7b-instruct-awq-4bit":
alias:
- 7b-4bit
project: vllm-chat
service_config:
name: qwen2
@@ -347,11 +313,9 @@
max_model_len: 2048
quantization: awq
extra_labels:
openllm_alias: 7b-4bit
openllm_alias: 7b-4bit,7b-instruct-4bit
openllm_hf_model_id: Qwen/Qwen2-7B-Instruct-AWQ
"qwen2:7b-instruct-fp16":
alias:
- 7b
project: vllm-chat
service_config:
name: qwen2
@@ -365,11 +329,9 @@
max_model_len: 2048
dtype: half
extra_labels:
openllm_alias: 7b
openllm_alias: 7b,7b-instruct
openllm_hf_model_id: Qwen/Qwen2-7B-Instruct
"qwen2:72b-instruct-awq-4bit":
alias:
- 72b-4bit
project: vllm-chat
service_config:
name: qwen2
@@ -383,11 +345,9 @@
max_model_len: 2048
quantization: awq
extra_labels:
openllm_alias: 72b-4bit
openllm_alias: 72b-4bit,72b-instruct-4bit
openllm_hf_model_id: Qwen/Qwen2-72B-Instruct-AWQ
"qwen2:57b-a14b-instruct-fp16":
alias:
- 57b-a14b
project: vllm-chat
service_config:
name: qwen2
@@ -401,11 +361,9 @@
max_model_len: 2048
dtype: half
extra_labels:
openllm_alias: 57b-a14b
openllm_alias: 57b-a14b,57b-a14b-instruct
openllm_hf_model_id: Qwen/Qwen2-57B-A14B-Instruct
"qwen2:72b-instruct-fp16":
alias:
- 72b
project: vllm-chat
service_config:
name: qwen2
@@ -419,5 +377,5 @@
max_model_len: 2048
dtype: half
extra_labels:
openllm_alias: 72b
openllm_alias: 72b,72b-instruct
openllm_hf_model_id: Qwen/Qwen2-72B-Instruct