NousResearch/atropos
1
0
Fork 0
mirror of https://github.com/NousResearch/atropos.git synced 2026-04-22 16:48:57 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions 1

major refactor 4

Browse source
This commit is contained in:
Jai Suphavadeeprasit 2026-01-22 13:39:34 -05:00
parent 1b1d8a5909
commit 59f337e34f
4 changed files with 18 additions and 9 deletions
Download patch file Download diff file Expand all files Collapse all files

9
example_trainer/scripts/run_comparison.sh
View file

@ -192,10 +192,11 @@ echo " Starting environment server..."
python -u environments/gsm8k_server.py serve \
--env.tokenizer_name "$MODEL" \
--env.use_wandb=False \
--env.rollout_server_url "http://localhost:8001" \
--openai.model_name "$MODEL" \
--openai.base_url "http://localhost:9001/v1" \
--openai.server_type vllm \
--server.port 8001 \
--slurm false \
> $LOGDIR/env_legacy.log 2>&1 &
LEGACY_ENV_PID=$!
echo " ✓ Environment server started (PID: $LEGACY_ENV_PID)"
@ -233,10 +234,11 @@ echo " Starting environment server..."
python -u environments/gsm8k_server.py serve \
--env.tokenizer_name "$MODEL" \
--env.use_wandb=False \
--env.rollout_server_url "http://localhost:8002" \
--openai.model_name "$MODEL" \
--openai.base_url "http://localhost:9002/v1" \
--openai.server_type vllm \
--server.port 8002 \
--slurm false \
> $LOGDIR/env_shared.log 2>&1 &
SHARED_ENV_PID=$!
echo " ✓ Environment server started (PID: $SHARED_ENV_PID)"
@ -293,10 +295,11 @@ echo " Starting environment server..."
python -u environments/gsm8k_server.py serve \
--env.tokenizer_name "$MODEL" \
--env.use_wandb=False \
--env.rollout_server_url "http://localhost:8003" \
--openai.model_name "$MODEL" \
--openai.base_url "http://localhost:9003/v1" \
--openai.server_type vllm \
--server.port 8003 \
--slurm false \
> $LOGDIR/env_lora.log 2>&1 &
LORA_ENV_PID=$!
echo " ✓ Environment server started (PID: $LORA_ENV_PID)"

6
example_trainer/scripts/run_concurrent_tests.sh
View file

@ -139,10 +139,11 @@ echo "[3/6] Starting LoRA GSM8k environment..."
python -u environments/gsm8k_server.py serve \
--env.tokenizer_name "$MODEL" \
--env.use_wandb=False \
--env.rollout_server_url "http://localhost:${LORA_GSM8K_PORT}" \
--openai.model_name "$MODEL" \
--openai.base_url "http://localhost:${LORA_VLLM_PORT}/v1" \
--openai.server_type vllm \
--server.port $LORA_GSM8K_PORT \
--slurm false \
> "${LOG_DIR}/lora_gsm8k.log" 2>&1 &
LORA_GSM8K_PID=$!
echo " PID: $LORA_GSM8K_PID"
@ -152,10 +153,11 @@ echo "[4/6] Starting Single-Copy GSM8k environment..."
python -u environments/gsm8k_server.py serve \
--env.tokenizer_name "$MODEL" \
--env.use_wandb=False \
--env.rollout_server_url "http://localhost:${SINGLE_COPY_GSM8K_PORT}" \
--openai.model_name "$MODEL" \
--openai.base_url "http://localhost:${SINGLE_COPY_VLLM_PORT}/v1" \
--openai.server_type vllm \
--server.port $SINGLE_COPY_GSM8K_PORT \
--slurm false \
> "${LOG_DIR}/single_copy_gsm8k.log" 2>&1 &
SINGLE_COPY_GSM8K_PID=$!
echo " PID: $SINGLE_COPY_GSM8K_PID"

3
example_trainer/scripts/test_single_copy_mode.sh
View file

@ -82,10 +82,11 @@ echo "[2/4] Starting GSM8k environment..."
python -u environments/gsm8k_server.py serve \
--env.tokenizer_name "$MODEL" \
--env.use_wandb=False \
--env.rollout_server_url "http://localhost:${GSM8K_PORT}" \
--openai.model_name "$MODEL" \
--openai.base_url "http://localhost:${VLLM_PORT}/v1" \
--openai.server_type vllm \
--server.port $GSM8K_PORT \
--slurm false \
> "${LOG_DIR}/gsm8k.log" 2>&1 &
echo "Waiting for GSM8k (10s)..."