mirror of
https://github.com/aljazceru/Auto-GPT.git
synced 2026-01-07 16:24:19 +01:00
Benchmark all agents (#5170)
This commit is contained in:
4
.github/workflows/benchmark-ci.yml
vendored
4
.github/workflows/benchmark-ci.yml
vendored
@@ -144,8 +144,10 @@ jobs:
|
||||
working-directory: ./benchmark/
|
||||
run: |
|
||||
mkdir agent
|
||||
link=$(jq -r '.["'"$AGENT_NAME"'"].url' agents_to_benchmark.json)
|
||||
branch=$(jq -r '.["'"$AGENT_NAME"'"].branch' agents_to_benchmark.json)
|
||||
cd agent
|
||||
git clone https://github.com/SilenNaihin/mini-agi -b benchmark-integration
|
||||
git clone "$link" -b "$branch"
|
||||
cd $AGENT_NAME
|
||||
prefix=""
|
||||
if [ "$AGENT_NAME" == "gpt-engineer" ]; then
|
||||
|
||||
38
benchmark/agents_to_benchmark.json
Normal file
38
benchmark/agents_to_benchmark.json
Normal file
@@ -0,0 +1,38 @@
|
||||
{
|
||||
"Auto-GPT": {
|
||||
"url": "https://github.com/Significant-Gravitas/Auto-GPT",
|
||||
"branch": "master"
|
||||
},
|
||||
"gpt-engineer": {
|
||||
"url": "https://github.com/merwanehamadi/gpt-engineer.git",
|
||||
"branch": "benchmark-integration"
|
||||
},
|
||||
"mini-agi": {
|
||||
"url": "https://github.com/SilenNaihin/mini-agi.git",
|
||||
"branch": "benchmark-integration"
|
||||
},
|
||||
"smol-developer": {
|
||||
"url": "https://github.com/e2b-dev/smol-developer.git",
|
||||
"branch": "benchmarks"
|
||||
},
|
||||
"SuperAGI": {
|
||||
"url": "https://github.com/SilenNaihin/SuperAGI.git",
|
||||
"branch": "benchmark-integration"
|
||||
},
|
||||
"BabyAGI": {
|
||||
"url": "https://github.com/SilenNaihin/babyagi.git",
|
||||
"branch": "benchmark-integration"
|
||||
},
|
||||
"beebot": {
|
||||
"url": "https://github.com/AutoPackAI/beebot.git",
|
||||
"branch": "main"
|
||||
},
|
||||
"PolyGPT": {
|
||||
"url": "https://github.com/polywrap/PolyGPT.git",
|
||||
"branch": "nerfzael-use-local-wrap-library"
|
||||
},
|
||||
"Turbo": {
|
||||
"url": "https://github.com/lc0rp/Auto-GPT-Turbo.git",
|
||||
"branch": "main"
|
||||
}
|
||||
}
|
||||
Reference in New Issue
Block a user