Merge pull request #426 from exo-explore/tinygrad_ci_test

Tinygrad ci test

.circleci/config.yml

@@ -20,6 +20,12 @@ commands:
           command: |
             source env/bin/activate
 
+            # Set CLANG=1 for tinygrad only
+            if [ "<<parameters.inference_engine>>" = "tinygrad" ]; then
+              pip install llvmlite
+              export TOKENIZERS_PARALLELISM=true SUPPORT_BF16=0 CLANG=1
+            fi
+
             # Start first instance
             HF_HOME="$(pwd)/.hf_cache_node1" DEBUG_DISCOVERY=7 DEBUG=7 exo --inference-engine <<parameters.inference_engine>> --node-id "node1" --listen-port 5678 --broadcast-port 5679 --chatgpt-api-port 8000 --chatgpt-api-response-timeout 900 2>&1 | tee output1.log &
             PID1=$!
@@ -48,13 +54,6 @@ commands:
             # Check processes before proceeding
             check_processes
 
-            # Special handling for dummy engine
-            if [ "<<parameters.inference_engine>>" = "dummy" ]; then
-              expected_content="This is a dummy response"
-            else
-              expected_content="Michael Jackson"
-            fi
-
             echo "Sending request to first instance..."
             response_1=$(curl -s http://localhost:8000/v1/chat/completions \
               -H "Content-Type: application/json" \
@@ -223,29 +222,29 @@ jobs:
       - checkout
       - run: system_profiler SPHardwareDataType
 
-  # chatgpt_api_integration_test_tinygrad:
-  #   macos:
-  #     xcode: "16.0.0"
-  #   resource_class: m2pro.large
-  #   steps:
-  #     - checkout
-  #     - run:
-  #         name: Set up Python
-  #         command: |
-  #           brew install python@3.12
-  #           python3.12 -m venv env
-  #           source env/bin/activate
-  #     - run:
-  #         name: Install dependencies
-  #         command: |
-  #           source env/bin/activate
-  #           pip install --upgrade pip
-  #           pip install .
-  #     - run_chatgpt_api_test:
-  #         inference_engine: tinygrad
-  #         model_id: llama-3-8b
-  #         prompt: "Keep responses concise. Who was the king of pop?"
-  #         expected_output: "Michael Jackson"
+  chatgpt_api_integration_test_tinygrad:
+    macos:
+      xcode: "16.0.0"
+    resource_class: m2pro.large
+    steps:
+      - checkout
+      - run:
+          name: Set up Python
+          command: |
+            brew install python@3.12
+            python3.12 -m venv env
+            source env/bin/activate
+      - run:
+          name: Install dependencies
+          command: |
+            source env/bin/activate
+            pip install --upgrade pip
+            pip install .
+      - run_chatgpt_api_test:
+          inference_engine: tinygrad
+          model_id: llama-3.2-1b
+          prompt: "Keep responses concise. Who was the king of pop?"
+          expected_output: "Michael Jackson"
 
 workflows:
   version: 2
@@ -254,6 +253,6 @@ workflows:
       - unit_test
       - discovery_integration_test
       - chatgpt_api_integration_test_mlx
+      - chatgpt_api_integration_test_tinygrad
       - chatgpt_api_integration_test_dummy
       - test_macos_m1
-      # - chatgpt_api_integration_test_tinygrad