Use `python -m petals.cli.run_server` instead of `python -m

cli.run_server`

.github/workflows/run-tests.yaml

@@ -37,7 +37,7 @@ jobs:
       - name: Convert model and push to hub
         run: |
           export HF_TAG=$(python -c "import os; print(os.environ.get('GITHUB_HEAD_REF') or os.environ.get('GITHUB_REF_NAME'))")
-          python -m cli.convert_model --model bigscience/bloom-560m  --output_path ./converted_model \
+          python -m petals.cli.convert_model --model bigscience/bloom-560m  --output_path ./converted_model \
             --output_repo bloom-testing/test-bloomd-560m-$HF_TAG --use_auth_token $BLOOM_TESTING_WRITE_TOKEN \
             --resize_token_embeddings 50000
 
@@ -70,7 +70,7 @@ jobs:
           export MODEL_NAME=bloom-testing/test-bloomd-560m-$HF_TAG
           export REF_NAME=bigscience/bloom-560m
 
-          python -m cli.run_server --converted_model_name_or_path $MODEL_NAME --block_indices 0:12 \
+          python -m petals.cli.run_server --converted_model_name_or_path $MODEL_NAME --block_indices 0:12 \
             --new_swarm --identity tests/test.id --host_maddrs /ip4/127.0.0.1/tcp/31337 --throughput 1 \
             --torch_dtype float32 --compression NONE --attn_cache_size 0.2GiB &> server1.log &
           SERVER1_PID=$!
@@ -80,21 +80,21 @@ jobs:
           export INITIAL_PEERS=/ip4/127.0.0.1/tcp/31337/p2p/QmS9KwZptnVdB9FFV7uGgaTq4sEKBwcYeKZDfSpyKDUd1g
           # ^-- server 1 multiaddr is determined by --identity and --host_maddrs
 
-          python -m cli.run_server --converted_model_name_or_path $MODEL_NAME --block_indices 12:22 \
+          python -m petals.cli.run_server --converted_model_name_or_path $MODEL_NAME --block_indices 12:22 \
             --initial_peers $INITIAL_PEERS --throughput 1 --torch_dtype float32 &> server2.log &
           SERVER2_PID=$!
 
           sleep 10 # wait for initial servers to declare blocks, then let server decide which blocks to serve
 
-          python -m cli.run_server --converted_model_name_or_path $MODEL_NAME --block_indices 0:6 \
+          python -m petals.cli.run_server --converted_model_name_or_path $MODEL_NAME --block_indices 0:6 \
             --initial_peers $INITIAL_PEERS --throughput 1 --torch_dtype float32 &> server3.log &
           SERVER3_PID=$!
 
-          python -m cli.run_server --converted_model_name_or_path $MODEL_NAME --block_indices 4:16 \
+          python -m petals.cli.run_server --converted_model_name_or_path $MODEL_NAME --block_indices 4:16 \
             --torch_dtype float32 --initial_peers $INITIAL_PEERS --throughput 1 &> server4.log &
           SERVER4_PID=$!
 
-          python -m cli.run_server --converted_model_name_or_path $MODEL_NAME --num_blocks 3 \
+          python -m petals.cli.run_server --converted_model_name_or_path $MODEL_NAME --num_blocks 3 \
             --initial_peers $INITIAL_PEERS --throughput 1 --torch_dtype float32 &> server5.log &
           SERVER5_PID=$!
 

README.md

@@ -107,7 +107,7 @@ For a detailed instruction with larger models, see ["Launch your own swarm"](htt
 
 First, run a couple of servers, each in a separate shell. To launch your first server, run:
 ```bash
-python -m cli.run_server bloom-testing/test-bloomd-560m-main --num_blocks 8 --torch_dtype float32 \
+python -m petals.cli.run_server bloom-testing/test-bloomd-560m-main --num_blocks 8 --torch_dtype float32 \
   --host_maddrs /ip4/127.0.0.1/tcp/31337   # use port 31337, local connections only
 ```
 
@@ -124,7 +124,7 @@ Mon Day 01:23:45.678 [INFO] Running DHT node on ['/ip4/127.0.0.1/tcp/31337/p2p/A
 You can use this address (`/ip4/whatever/else`) to connect additional servers. Open another terminal and run:
 
 ```bash
-python -m cli.run_server bloom-testing/test-bloomd-560m-main --num_blocks 8 --torch_dtype float32 \
+python -m petals.cli.run_server bloom-testing/test-bloomd-560m-main --num_blocks 8 --torch_dtype float32 \
   --host_maddrs /ip4/127.0.0.1/tcp/0 \
   --initial_peers /ip4/127.0... # <-- TODO: Copy the address of another server here
 # e.g. --initial_peers /ip4/127.0.0.1/tcp/31337/p2p/QmS1GecIfYouAreReadingThisYouNeedToCopyYourServerAddressCBBq
@@ -176,14 +176,14 @@ To run minimalistic tests, spin up some servers:
 ```bash
 export MODEL_NAME=bloom-testing/test-bloomd-560m-main
 export INITIAL_PEERS=/ip4/127.0.0.1/tcp/31337/p2p/QmS9KwZptnVdB9FFV7uGgaTq4sEKBwcYeKZDfSpyKDUd1g
-python -m cli.run_server $MODEL_NAME --block_indices 0:12 --throughput 1 --torch_dtype float32 \
+python -m petals.cli.run_server $MODEL_NAME --block_indices 0:12 --throughput 1 --torch_dtype float32 \
   --identity tests/test.id --host_maddrs /ip4/127.0.0.1/tcp/31337  &> server1.log &
 sleep 5  # wait for the first server to initialize DHT
-python -m cli.run_server $MODEL_NAME --block_indices 12:24 --throughput 1 --torch_dtype float32 \
+python -m petals.cli.run_server $MODEL_NAME --block_indices 12:24 --throughput 1 --torch_dtype float32 \
   --initial_peers /ip4/127.0.0.1/tcp/31337/p2p/QmS9KwZptnVdB9FFV7uGgaTq4sEKBwcYeKZDfSpyKDUd1g &> server2.log &
 
 tail -f server1.log server2.log  # view logs for both servers
-# after you're done, kill servers with 'pkill -f cli.run_server'
+# after you're done, kill servers with 'pkill -f petals.cli.run_server'
 ```
 
 Then launch pytest: