WIP: adding workflows & template

.github/workflows/release_model.yaml

@@ -0,0 +1,37 @@
+name: release_model
+
+env:
+  model_name: "replicate/llama-7b"
+
+on:
+  push:
+    branches:
+      - main
+
+jobs:
+  release:
+    runs-on: [self-hosted, gpu-a100]
+    strategy: 
+      matrix:
+        model: ['llama-7b']
+    steps:
+      - uses: actions/checkout@v3
+
+      - name: Install Cog
+        run: |
+          sudo curl -o /usr/local/bin/cog -L https://github.com/replicate/cog/releases/latest/download/cog_`uname -s`_`uname -m`
+          sudo chmod +x /usr/local/bin/cog
+
+      - name: Select model
+        run: 
+          cog run python select_model.py --model_name ${{ matrix.model }} --model_path ${{ secrets.LLAMA_7B_PATH }}
+
+      - name: Log in to Replicate
+        env:
+          REPLICATE_API_TOKEN: ${{ secrets.REPLICATE_API_TOKEN }}
+        run: |
+          echo $REPLICATE_API_TOKEN | cog login --token-stdin
+
+      - name: Push to Replicate
+        run: |
+          cog push r8.im/${{ env.model_name }}

.github/workflows/run_tests.yaml

@@ -0,0 +1,33 @@
+name: run_tests
+
+on:
+  pull_request:
+    branches:
+      - main
+
+jobs:
+  test:
+    runs-on: [self-hosted, gpu-a100]
+    strategy: 
+      matrix:
+        model: ['llama-7b']
+    steps:
+      - uses: actions/checkout@v3
+      - uses: actions/setup-python@v3
+        with:
+          python-version: 3.8
+
+      - name: Install Python dependencies
+        run: pip install pytest
+
+      - name: Install Cog
+        run: |
+          sudo curl -o /usr/local/bin/cog -L https://github.com/replicate/cog/releases/latest/download/cog_`uname -s`_`uname -m`
+          sudo chmod +x /usr/local/bin/cog
+
+      - name: Select model
+        run: 
+          cog run python select_model.py --model_name ${{ matrix.model }} --model_path ${{ secrets.LLAMA_7B_PATH }}
+
+      - name: Run tests
+        run: pytest -vv test/

select_model.py

@@ -7,11 +7,11 @@
 CONFIGS = {
     "llama-7b": {
         "cog_yaml_parameters": {"predictor":"predict.py:Predictor"},
-        "config_py_parameters": {"model_name": "SET_ME", "config_location": "llama_weights/llama-7b"}
+        "config_py_parameters": {"model_path": "SET_ME", "config_location": "llama_weights/llama-7b"}
     },
     "llama-13b": {
         "cog_yaml_parameters": {"predictor":"predict.py:Predictor"},
-        "config_py_parameters": {"model_name": "SET_ME", "config_location": "llama_weights/llama-13b"}
+        "config_py_parameters": {"model_path": "SET_ME", "config_location": "llama_weights/llama-13b"}
     },
 }
 
@@ -36,15 +36,17 @@ def write_one_config(template_fpath: str, fname_out: str, config: dict):
     os.chmod(fname_out, new_permissions)
 
 
-def write_configs(model_name):
+def write_configs(model_name, model_path):
     master_config = CONFIGS[model_name]
-    #write_one_config("templates/cog_template.yaml", "cog.yaml", master_config['cog_yaml_parameters'])
-    write_one_config("templates/config_template.py", "cronfig.py", master_config['config_py_parameters'])
+    cfg = master_config["config_py_parameters"]    
+    cfg['model_path'] = model_path
+    write_one_config("templates/config_template.py", "config.py", master_config['config_py_parameters'])
 
 
 if __name__ == '__main__':
     parser = ArgumentParser()
     parser.add_argument("--model_name", default="llama-7b", help="name of the flan-t5 model you want to configure cog for")
+    parser.add_argument("--model_path", default="llama-7b", help="path to llama model, in cloud storage or locally")
     args = parser.parse_args()
 
     write_configs(args.model_name)

templates/config_template.py

@@ -11,7 +11,7 @@
 
 from subclass import YieldingLlama
 
-DEFAULT_MODEL_NAME = "{{model_name}}" # path from which we pull weights when there's no COG_WEIGHTS environment variable
+DEFAULT_MODEL_NAME = "{{model_path}}" # path from which we pull weights when there's no COG_WEIGHTS environment variable
 TOKENIZER_NAME = "llama_weights/tokenizer"
 CONFIG_LOCATION = "{{config_location}}"
 
@@ -60,7 +60,7 @@ def load_tensorizer(
     weights = str(weights)
     local_weights = "/src/llama_tensors"
     print("Deserializing weights...")
-    if 'http' in weights:
+    if 'http' in weights or 'gs' in weights:
         pull_gcp_file(weights, local_weights)
     else:
         local_weights = weights