chore!: rename agents -> vlagents

Author: juelg

.gitignore

@@ -12,3 +12,4 @@ wandb
 run_logs
 .ipynb_checkpoints
 .not*
+dist

.vscode/launch.json

@@ -12,13 +12,13 @@
             "console": "integratedTerminal",
             "args": ["-vv"],
             "justMyCode": true,
-            "cwd": "${workspaceFolder}"
+            "cwd": "${workspaceFolder}/src"
         },
         {
             "name": "Debug server with test policy",
             "type": "debugpy",
             "request": "launch",
-            "module": "agents",
+            "module": "vlagents",
             "console": "integratedTerminal",
             "args": ["start-server", "test", "--port=8080", "--host=localhost", "--kwargs={}"],
             "justMyCode": true,

README.md

@@ -1,5 +1,5 @@
-# Agents
-Agents is a python library that allows to separate next action prediction from policy networks from action execution in simulated or real environments.
+# VLAgents
+VLAgents is a python library that allows to separate next action prediction from policy networks from action execution in simulated or real environments.
 It defines an interface for policies and for environments.
 The policies run independent in their own virtual environment, potentially on a different computer, and can be queried for an action (in principle similar to the chatgpt api).
 
@@ -15,18 +15,18 @@ The work also includes a section on related engineering challenges regarding jax
 
 ### Local Installation
 ```shell
-git clone https://github.com/juelg/agents.git
-cd agents
+git clone https://github.com/juelg/vlagents.git
+cd vlagents
 pip install -ve .
 ```
 
 ### Repo Installation
 ```shell
-pip install git+https://github.com/juelg/agents.git
+pip install git+https://github.com/juelg/vlagents.git
 ```
 
 ### Environment and Policy Installation
-On top of agents you can then install a simulation environment where the agent acts.
+On top of vlagents you can then install a simulation environment where the agent acts.
 We currently support [maniskill](https://github.com/haosulab/ManiSkill) with more to come.
 In order to avoid dependency conflicts, use a second conda/pip environment to install your policy.
 We currently support [octo](https://github.com/octo-models/octo) and [openvla](https://github.com/openvla/openvla).
@@ -40,7 +40,7 @@ conda install nvidia/label/cuda-11.8.0::cuda --no-channel-priority
 conda install conda-forge::cudnn=8.9
 # octo dependencies
 pip install git+https://github.com/octo-models/octo.git@241fb3514b7c40957a86d869fecb7c7fc353f540
-pip install -r agents/utils/fixed_octo_requirements.txt
+pip install -r vlagents/utils/fixed_octo_requirements.txt
 # for gpu support:
 pip install --upgrade "jax[cuda11_pip]==0.4.20" -f https://storage.googleapis.com/jax-releases/jax_cuda_releases.html
 ```
@@ -54,9 +54,9 @@ print(xla_bridge.get_backend().platform)
 ```
 
 
-Install the agents library on top:
+Install the vlagents library on top:
 ```shell
-pip install git+https://github.com/juelg/agents.git
+pip install git+https://github.com/juelg/vlagents.git
 ```
 
 For more details, see the [Octo github page](https://github.com/octo-models/octo).
@@ -93,9 +93,9 @@ Install OpenVLA
 pip install git+https://github.com/openvla/openvla.git@46b752f477cc5773cc1234b2e82c0e2130e4e890
 ```
 
-Install the agents library on top:
+Install the vlagents library on top:
 ```shell
-pip install git+https://github.com/juelg/agents.git
+pip install git+https://github.com/juelg/vlagents.git
 ```
 
 For more details, see the [OpenVLA github page](https://github.com/openvla/openvla).
@@ -129,25 +129,25 @@ git clone git@github.com:facebookresearch/vjepa2.git
 cd vjepa2
 pip install -e .
 
-pip install git+https://github.com/juelg/agents.git
+pip install git+https://github.com/juelg/vlagents.git
 pip install -ve .
 
 ```
 
 ## Usage
-To start an agents server use the `start-server` command where `kwargs` is a dictionary of the constructor arguments of the policy you want to start e.g.
+To start an vlagents server use the `start-server` command where `kwargs` is a dictionary of the constructor arguments of the policy you want to start e.g.
 ```shell
 # octo
-python -m agents start-server octo --host localhost --port 8080 --kwargs '{"checkpoint_path": "hf://Juelg/octo-base-1.5-finetuned-maniskill", "checkpoint_step": None, "horizon": 1, "unnorm_key": []}'
+python -m vlagents start-server octo --host localhost --port 8080 --kwargs '{"checkpoint_path": "hf://Juelg/octo-base-1.5-finetuned-maniskill", "checkpoint_step": None, "horizon": 1, "unnorm_key": []}'
 
 # openvla
-python -m agents start-server openvla --host localhost --port 8080 --kwargs '{"checkpoint_path": "Juelg/openvla-7b-finetuned-maniskill", "device": "cuda:0", "attn_implementation": "flash_attention_2", "unnorm_key": "maniskill_human:7.0.0", "checkpoint_step": 40000}'
+python -m vlagents start-server openvla --host localhost --port 8080 --kwargs '{"checkpoint_path": "Juelg/openvla-7b-finetuned-maniskill", "device": "cuda:0", "attn_implementation": "flash_attention_2", "unnorm_key": "maniskill_human:7.0.0", "checkpoint_step": 40000}'
 
 # openpi
-python -m agents start-server openpi --port=8080 --host=localhost --kwargs='{"checkpoint_path": "<path to checkpoint>/{checkpoint_step}", "model_name": "pi0_rcs", "checkpoint_step": <checkpoint_step>}' # leave "{checkpoint_step}" it will be replaced, "model_name" is the key for the training config
+python -m vlagents start-server openpi --port=8080 --host=localhost --kwargs='{"checkpoint_path": "<path to checkpoint>/{checkpoint_step}", "model_name": "pi0_rcs", "checkpoint_step": <checkpoint_step>}' # leave "{checkpoint_step}" it will be replaced, "model_name" is the key for the training config
 
 # vjepa2-ac
-python -m agents start-server vjepa --port=20997 --host=0.0.0.0 --kwargs='{"cfg_path": "configs/inference/vjepa2-ac-vitg/<your_config>.yaml", "model_name": "vjepa2_ac_vit_giant", "default_checkpoint_path": "../.cache/torch/hub/checkpoints/vjepa2-ac-vitg.pt"}'
+python -m vlagents start-server vjepa --port=20997 --host=0.0.0.0 --kwargs='{"cfg_path": "configs/inference/vjepa2-ac-vitg/<your_config>.yaml", "model_name": "vjepa2_ac_vit_giant", "default_checkpoint_path": "../.cache/torch/hub/checkpoints/vjepa2-ac-vitg.pt"}'
 ```
 
 
@@ -159,12 +159,12 @@ In both cases environment and arguments as well as policy and arguments and wand
 ## Contribution
 
 ### New Policy
-In order to extend the library with a new policy network, extend the `Agent` class in [policies.py](src/agents/policies.py).
+In order to extend the library with a new policy network, extend the `Agent` class in [policies.py](src/vlagents/policies.py).
 It is important to only invoke policy specific imports in the class functions, as each policy can have its own dependencies.
 
 
 ### New Environment
-In order to extend the library with a new agent environment, extend the `EvaluatorEnv` class in [evaluator_envs.py](src/agents/evaluator_envs.py).
+In order to extend the library with a new agent environment, extend the `EvaluatorEnv` class in [evaluator_envs.py](src/vlagents/evaluator_envs.py).
 
 
 ### Developer Tools

pyproject.toml

@@ -3,7 +3,7 @@ requires = ["setuptools"]
 build-backend = "setuptools.build_meta"
 
 [project]
-name = "agents"
+name = "vlagents"
 version = "0.0.1"
 description = "Modular VLA and Environment interfaces."
 dependencies = [

src/tests/test_connection.py

@@ -3,9 +3,9 @@
 
 import numpy as np
 
-from agents.client import RemoteAgent
-from agents.evaluator_envs import start_server
-from agents.policies import Act, Obs
+from vlagents.client import RemoteAgent
+from vlagents.evaluator_envs import start_server
+from vlagents.policies import Act, Obs
 
 
 def _test_connection(agent: RemoteAgent):

src/agents/__init__.py src/vlagents/__init__.pysrc/agents/__init__.py renamed to src/vlagents/__init__.py

@@ -15,11 +15,11 @@
 # when started from jupyter notebook
 os.environ["MPLBACKEND"] = "Agg"
 
-from agents.evaluator_envs import AgentConfig, EvalConfig, evaluation, write_results
-from agents.policies import AGENTS
-from agents.server import AgentService
+from vlagents.evaluator_envs import AgentConfig, EvalConfig, evaluation, write_results
+from vlagents.policies import AGENTS
+from vlagents.server import AgentService
 
-main_app = typer.Typer(help="CLI tool for the agents library.")
+main_app = typer.Typer(help="CLI tool for the vlagents library.")
 
 
 def wandb_log_git_diff(path: str):

src/agents/__main__.py src/vlagents/__main__.pysrc/agents/__main__.py renamed to src/vlagents/__main__.py

@@ -9,7 +9,7 @@
 import rpyc
 import simplejpeg
 
-from agents.policies import Act, Agent, CameraDataType, Obs, SharedMemoryPayload
+from vlagents.policies import Act, Agent, CameraDataType, Obs, SharedMemoryPayload
 
 
 def dataclass_from_dict(klass, d):

src/agents/client.py src/vlagents/client.pysrc/agents/client.py renamed to src/vlagents/client.py

@@ -18,9 +18,9 @@
 from simple_slurm import Slurm
 from tqdm import tqdm
 
-from agents.client import RemoteAgent
-from agents.policies import Act, Agent, Obs
-from agents.wrappers import HumanCameraWrapper
+from vlagents.client import RemoteAgent
+from vlagents.policies import Act, Agent, Obs
+from vlagents.wrappers import HumanCameraWrapper
 
 logging.basicConfig(
     format="%(asctime)s,%(msecs)d %(name)s %(levelname)s %(message)s",
@@ -339,7 +339,7 @@ def start_server(
     cmd = [
         python_path.format(agent_name=agent_name),
         "-m",
-        "agents",
+        "vlagents",
         "start-server",
         f"{agent_name}",
         f"--port={port}",
@@ -415,7 +415,7 @@ def run_eval_during_training(
         cmd = ["python"]
     cmd += [
         "-m",
-        "agents" "run-eval-during-training",
+        "vlagents" "run-eval-during-training",
         f"--agent-cfg={json.dumps(asdict(agent_cfg))}" f"--episodes={episodes}",
         f"--n-processes={n_processes}",
         f"--eval-cfgs={json.dumps([asdict(cfg) for cfg in eval_cfgs])}",
@@ -452,7 +452,7 @@ def run_eval_post_training(
         shlex.join(
             [
                 "-m",
-                "agents",
+                "vlagents",
                 "run-eval-post-training",
                 f"--agent-cfg={json.dumps(asdict(agent_cfg))}",
                 f"--episodes={episodes}",

src/agents/evaluator_envs.py src/vlagents/evaluator_envs.pysrc/agents/evaluator_envs.py renamed to src/vlagents/evaluator_envs.py

@@ -343,7 +343,7 @@ def act(self, obs: Obs) -> Act:
         )
         action_chunk = self.policy.infer(observation)["actions"]
 
-        # convert gripper action into agents format
+        # convert gripper action into vlagents format
         action_chunk[:, -1] = 1 - action_chunk[:, -1]
         self._cached_action_chunk = action_chunk