← back to Exo
use llama 3.1 in tests
67ad3f57a1c6804d11fc446500d67fc4aecbdd3e · 2024-08-02 10:56:29 +0100 · Alex Cheema
Files touched
Diff
commit 67ad3f57a1c6804d11fc446500d67fc4aecbdd3e
Author: Alex Cheema <alexcheema123@gmail.com>
Date: Fri Aug 2 10:56:29 2024 +0100
use llama 3.1 in tests
---
.circleci/config.yml | 10 +++++-----
1 file changed, 5 insertions(+), 5 deletions(-)
diff --git a/.circleci/config.yml b/.circleci/config.yml
index b78abf87..db2aa82c 100644
--- a/.circleci/config.yml
+++ b/.circleci/config.yml
@@ -32,7 +32,7 @@ jobs:
name: Run tests
command: |
source env/bin/activate
- ls ~/.cache/huggingface/hub/models--mlx-community--Meta-Llama-3-8B-Instruct-4bit/**/* || true
+ ls ~/.cache/huggingface/hub/models--mlx-community--Meta-Llama-3.1-8B-Instruct-4bit/**/* || true
METAL_XCODE=1 python3 -m exo.inference.test_inference_engine
- save_cache:
paths:
@@ -112,7 +112,7 @@ jobs:
command: |
source env/bin/activate
# Check if cached files are present
- ls ~/.cache/huggingface/hub/models--mlx-community--Meta-Llama-3-8B-Instruct-4bit/**/* || true
+ ls ~/.cache/huggingface/hub/models--mlx-community--Meta-Llama-3.1-8B-Instruct-4bit/**/* || true
# Start first instance
DEBUG_DISCOVERY=9 DEBUG=9 python3 main.py --inference-engine mlx --node-id "node1" --listen-port 5678 --broadcast-port 5679 --chatgpt-api-port 8000 --chatgpt-api-response-timeout-secs 900 > output1.log 2>&1 &
@@ -146,7 +146,7 @@ jobs:
curl -s http://localhost:8000/v1/chat/completions \
-H "Content-Type: application/json" \
-d '{
- "model": "llama-3-8b",
+ "model": "llama-3.1-8b",
"messages": [{"role": "user", "content": "Keep responses concise. Placeholder to load model..."}],
"temperature": 0.7
}'
@@ -157,7 +157,7 @@ jobs:
response_1=$(curl -s http://localhost:8000/v1/chat/completions \
-H "Content-Type: application/json" \
-d '{
- "model": "llama-3-8b",
+ "model": "llama-3.1-8b",
"messages": [{"role": "user", "content": "Keep responses concise. Who was the king of pop?"}],
"temperature": 0.7
}')
@@ -169,7 +169,7 @@ jobs:
response_2=$(curl -s http://localhost:8000/v1/chat/completions \
-H "Content-Type: application/json" \
-d '{
- "model": "llama-3-8b",
+ "model": "llama-3.1-8b",
"messages": [{"role": "user", "content": "Keep responses concise. Who was the king of pop?"}],
"temperature": 0.7
}')
← 92b66e27 fix cache
·
back to Exo
·
only remove the matching prefix from the prompt if its lengt 54ea5dbb →