evalmc 0.0.2__tar.gz → 0.0.4__tar.gz

evalmc-0.0.4/PKG-INFO

@@ -0,0 +1,136 @@
+Metadata-Version: 2.4
+Name: evalmc
+Version: 0.0.4
+Summary: MyoChallenge CLI — submit ONNX and Docker policies trained with MyoSuite 3
+Author-email: MyoChallenge Team <caggiano@gmail.com>
+License-Expression: Apache-2.0
+Project-URL: Homepage, https://mc26.myosuite.org
+Project-URL: MyoSuite, https://www.myosuite.org
+Project-URL: API, https://api.mc.myosuite.org
+Classifier: Development Status :: 4 - Beta
+Classifier: Intended Audience :: Science/Research
+Classifier: Operating System :: OS Independent
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Programming Language :: Python :: 3.13
+Classifier: Programming Language :: Python :: 3.14
+Classifier: Topic :: Scientific/Engineering
+Requires-Python: >=3.10
+Description-Content-Type: text/markdown
+License-File: LICENSE
+Requires-Dist: click>=8.0.0
+Requires-Dist: docker>=6.0.0
+Requires-Dist: requests>=2.25.0
+Requires-Dist: boto3>=1.26.0
+Dynamic: license-file
+
+# evalmc
+
+Submission CLI for [MyoChallenge](https://mc26.myosuite.org). Train with
+[MyoSuite 3](https://pypi.org/project/myosuite/) (`pip install -U myosuite`),
+then upload an ONNX graph or a Docker agent image.
+
+[![PyPI](https://img.shields.io/pypi/v/evalmc.svg)](https://pypi.org/project/evalmc/)
+[![Python](https://img.shields.io/pypi/pyversions/evalmc.svg)](https://pypi.org/project/evalmc/)
+[![License](https://img.shields.io/pypi/l/evalmc.svg)](https://pypi.org/project/evalmc/)
+
+Python **3.10–3.14**, same range as MyoSuite 3. Default API:
+`https://api.mc.myosuite.org`.
+
+```bash
+pip install -U evalmc myosuite
+```
+
+---
+
+## Train locally (MyoSuite 3)
+
+2026 ChaseTag uses the steered full-body 1v1 env on the Gymnasium API:
+
+```python
+from myosuite import make_env
+
+env = make_env("myoChallengeChaseTagFBVs-v0")
+obs, info = env.reset(seed=0)
+```
+
+| | |
+|---|---|
+| **Env** | `myoChallengeChaseTagFBVs-v0` |
+| **Task / phase** | `chasetag` / `1` |
+| **Observation** | `(546,)` — you are `agent_0`; last 2 dims are role one-hot |
+| **Action** | `(354,)` — muscle activations in `[0, 1]` |
+| **Scoring** | Title cycle: mid-cycle vs a fixed champion, then single-elim finals |
+
+CPU training (Stable-Baselines3) or GPU (`make_env(..., backend="mjlab")`) both
+export to the same ONNX contract. See [MyoSuite 3](https://www.myosuite.org).
+
+---
+
+## Quick start
+
+```bash
+evalmc host --set-host https://api.mc.myosuite.org
+evalmc login                          # email + token from the dashboard
+evalmc status
+evalmc submit-onnx model.onnx -t chasetag -p 1
+```
+
+Credentials are stored in `~/.myochallenge/`.
+
+---
+
+## Commands
+
+| Command | Purpose |
+|---|---|
+| `evalmc host` | Show or set the API host |
+| `evalmc login` | Save email + auth token |
+| `evalmc set_token` | Set the token non-interactively |
+| `evalmc status` | Confirm host and team |
+| `evalmc submit-onnx` | Upload a standalone ONNX policy |
+| `evalmc nominate` | Pick a scored artifact for title-cycle finals |
+| `evalmc push` | Submit a Docker / ECR agent image |
+
+### ONNX (recommended)
+
+Graph is `obs → action`. No `policy.py` required.
+
+```bash
+evalmc submit-onnx model.onnx -t chasetag -p 1
+```
+
+ChaseTag tensors: input `obs` `[batch, 546]`, output `action` `[batch, 354]`.
+
+### Title-cycle nomination
+
+Before freeze, choose which scored in-window submission plays in finals.
+Without a nomination the platform uses your best score.
+
+```bash
+evalmc nominate -t chasetag -p 1 --submission-id 123
+```
+
+### Docker
+
+For custom Python around the policy. Ship `policy.py` and/or `model.onnx`
+under `/policy` (or `/app`):
+
+```python
+class Policy:
+    def reset(self):
+        ...
+    def get_action(self, obs):
+        ...  # 1-D float array, length 354 for ChaseTag
+```
+
+```bash
+evalmc push my-agent:tag -t chasetag -p 1
+```
+
+---
+
+ChaseTag 2026 is invitation-only. Registration and submission are rejected until
+the team is on the whitelist. Leaderboards stay public.

evalmc-0.0.4/README.md

@@ -0,0 +1,108 @@
+# evalmc
+
+Submission CLI for [MyoChallenge](https://mc26.myosuite.org). Train with
+[MyoSuite 3](https://pypi.org/project/myosuite/) (`pip install -U myosuite`),
+then upload an ONNX graph or a Docker agent image.
+
+[![PyPI](https://img.shields.io/pypi/v/evalmc.svg)](https://pypi.org/project/evalmc/)
+[![Python](https://img.shields.io/pypi/pyversions/evalmc.svg)](https://pypi.org/project/evalmc/)
+[![License](https://img.shields.io/pypi/l/evalmc.svg)](https://pypi.org/project/evalmc/)
+
+Python **3.10–3.14**, same range as MyoSuite 3. Default API:
+`https://api.mc.myosuite.org`.
+
+```bash
+pip install -U evalmc myosuite
+```
+
+---
+
+## Train locally (MyoSuite 3)
+
+2026 ChaseTag uses the steered full-body 1v1 env on the Gymnasium API:
+
+```python
+from myosuite import make_env
+
+env = make_env("myoChallengeChaseTagFBVs-v0")
+obs, info = env.reset(seed=0)
+```
+
+| | |
+|---|---|
+| **Env** | `myoChallengeChaseTagFBVs-v0` |
+| **Task / phase** | `chasetag` / `1` |
+| **Observation** | `(546,)` — you are `agent_0`; last 2 dims are role one-hot |
+| **Action** | `(354,)` — muscle activations in `[0, 1]` |
+| **Scoring** | Title cycle: mid-cycle vs a fixed champion, then single-elim finals |
+
+CPU training (Stable-Baselines3) or GPU (`make_env(..., backend="mjlab")`) both
+export to the same ONNX contract. See [MyoSuite 3](https://www.myosuite.org).
+
+---
+
+## Quick start
+
+```bash
+evalmc host --set-host https://api.mc.myosuite.org
+evalmc login                          # email + token from the dashboard
+evalmc status
+evalmc submit-onnx model.onnx -t chasetag -p 1
+```
+
+Credentials are stored in `~/.myochallenge/`.
+
+---
+
+## Commands
+
+| Command | Purpose |
+|---|---|
+| `evalmc host` | Show or set the API host |
+| `evalmc login` | Save email + auth token |
+| `evalmc set_token` | Set the token non-interactively |
+| `evalmc status` | Confirm host and team |
+| `evalmc submit-onnx` | Upload a standalone ONNX policy |
+| `evalmc nominate` | Pick a scored artifact for title-cycle finals |
+| `evalmc push` | Submit a Docker / ECR agent image |
+
+### ONNX (recommended)
+
+Graph is `obs → action`. No `policy.py` required.
+
+```bash
+evalmc submit-onnx model.onnx -t chasetag -p 1
+```
+
+ChaseTag tensors: input `obs` `[batch, 546]`, output `action` `[batch, 354]`.
+
+### Title-cycle nomination
+
+Before freeze, choose which scored in-window submission plays in finals.
+Without a nomination the platform uses your best score.
+
+```bash
+evalmc nominate -t chasetag -p 1 --submission-id 123
+```
+
+### Docker
+
+For custom Python around the policy. Ship `policy.py` and/or `model.onnx`
+under `/policy` (or `/app`):
+
+```python
+class Policy:
+    def reset(self):
+        ...
+    def get_action(self, obs):
+        ...  # 1-D float array, length 354 for ChaseTag
+```
+
+```bash
+evalmc push my-agent:tag -t chasetag -p 1
+```
+
+---
+
+ChaseTag 2026 is invitation-only. Registration and submission are rejected until
+the team is on the whitelist. Leaderboards stay public.

{evalmc-0.0.2 → evalmc-0.0.4}/evalmc/main.py

@@ -18,13 +18,13 @@ from .nominate import nominate
 @click.version_option()
 def main():
     """
-    MyoChallenge submission CLI.
+    MyoChallenge submission CLI (MyoSuite 3).
 
     Two submission paths are available:
 
     \b
-    evalmc submit-onnx model.onnx -p 1 -t elbowflexion  # ONNX (elbow / ChaseTag)
-    evalmc push   my-image:tag -p 1 -t tabletennis  # Docker/ECR image push
+    evalmc submit-onnx model.onnx -p 1 -t chasetag   # ONNX (recommended)
+    evalmc push   my-image:tag -p 1 -t chasetag      # Docker / ECR
     """
     pass
 

{evalmc-0.0.2 → evalmc-0.0.4}/evalmc/submission.py

@@ -102,7 +102,7 @@ def _run_subprocess(cmd: list[str], step: str) -> None:
 @click.option(
     "-t",
     "--task",
-    help="challenge task name (e.g.,tabletennis, soccer)",
+    help="challenge task name (e.g. chasetag)",
     required=True,
 )
 @click.option(

{evalmc-0.0.2 → evalmc-0.0.4}/evalmc/submit_onnx.py

@@ -12,16 +12,16 @@ from .push_helpers import echo_submission_result, post_submission_file, resolve_
 @click.command("submit-onnx")
 @click.argument("onnx_path", type=click.Path(exists=True, path_type=Path))
 @click.option("-p", "--phase", required=True, type=int, help="Challenge phase number")
-@click.option("-t", "--task", required=True, help="Challenge task name (e.g. elbowflexion)")
+@click.option("-t", "--task", required=True, help="Challenge task name (e.g. chasetag)")
 @click.option("-e", "--email", default=None, help="Team email")
 @click.option("-a", "--auth-token", default=None, help="Auth token")
 def submit_onnx(onnx_path, phase, task, email, auth_token):
     """Submit a standalone ONNX model (no policy.py required).
 
-    Works for elbow flexion and ChaseTag (obs→action graphs).
+    ChaseTag (MyoSuite 3): obs [batch, 546] → action [batch, 354]
+    on myoChallengeChaseTagFBVs-v0.
 
     Example:
-      evalmc submit-onnx my_model.onnx -t elbowflexion -p 1
       evalmc submit-onnx chase.onnx -t chasetag -p 1
     """
     if onnx_path.suffix.lower() != ".onnx":

evalmc-0.0.4/evalmc.egg-info/PKG-INFO

@@ -0,0 +1,136 @@
+Metadata-Version: 2.4
+Name: evalmc
+Version: 0.0.4
+Summary: MyoChallenge CLI — submit ONNX and Docker policies trained with MyoSuite 3
+Author-email: MyoChallenge Team <caggiano@gmail.com>
+License-Expression: Apache-2.0
+Project-URL: Homepage, https://mc26.myosuite.org
+Project-URL: MyoSuite, https://www.myosuite.org
+Project-URL: API, https://api.mc.myosuite.org
+Classifier: Development Status :: 4 - Beta
+Classifier: Intended Audience :: Science/Research
+Classifier: Operating System :: OS Independent
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Programming Language :: Python :: 3.13
+Classifier: Programming Language :: Python :: 3.14
+Classifier: Topic :: Scientific/Engineering
+Requires-Python: >=3.10
+Description-Content-Type: text/markdown
+License-File: LICENSE
+Requires-Dist: click>=8.0.0
+Requires-Dist: docker>=6.0.0
+Requires-Dist: requests>=2.25.0
+Requires-Dist: boto3>=1.26.0
+Dynamic: license-file
+
+# evalmc
+
+Submission CLI for [MyoChallenge](https://mc26.myosuite.org). Train with
+[MyoSuite 3](https://pypi.org/project/myosuite/) (`pip install -U myosuite`),
+then upload an ONNX graph or a Docker agent image.
+
+[![PyPI](https://img.shields.io/pypi/v/evalmc.svg)](https://pypi.org/project/evalmc/)
+[![Python](https://img.shields.io/pypi/pyversions/evalmc.svg)](https://pypi.org/project/evalmc/)
+[![License](https://img.shields.io/pypi/l/evalmc.svg)](https://pypi.org/project/evalmc/)
+
+Python **3.10–3.14**, same range as MyoSuite 3. Default API:
+`https://api.mc.myosuite.org`.
+
+```bash
+pip install -U evalmc myosuite
+```
+
+---
+
+## Train locally (MyoSuite 3)
+
+2026 ChaseTag uses the steered full-body 1v1 env on the Gymnasium API:
+
+```python
+from myosuite import make_env
+
+env = make_env("myoChallengeChaseTagFBVs-v0")
+obs, info = env.reset(seed=0)
+```
+
+| | |
+|---|---|
+| **Env** | `myoChallengeChaseTagFBVs-v0` |
+| **Task / phase** | `chasetag` / `1` |
+| **Observation** | `(546,)` — you are `agent_0`; last 2 dims are role one-hot |
+| **Action** | `(354,)` — muscle activations in `[0, 1]` |
+| **Scoring** | Title cycle: mid-cycle vs a fixed champion, then single-elim finals |
+
+CPU training (Stable-Baselines3) or GPU (`make_env(..., backend="mjlab")`) both
+export to the same ONNX contract. See [MyoSuite 3](https://www.myosuite.org).
+
+---
+
+## Quick start
+
+```bash
+evalmc host --set-host https://api.mc.myosuite.org
+evalmc login                          # email + token from the dashboard
+evalmc status
+evalmc submit-onnx model.onnx -t chasetag -p 1
+```
+
+Credentials are stored in `~/.myochallenge/`.
+
+---
+
+## Commands
+
+| Command | Purpose |
+|---|---|
+| `evalmc host` | Show or set the API host |
+| `evalmc login` | Save email + auth token |
+| `evalmc set_token` | Set the token non-interactively |
+| `evalmc status` | Confirm host and team |
+| `evalmc submit-onnx` | Upload a standalone ONNX policy |
+| `evalmc nominate` | Pick a scored artifact for title-cycle finals |
+| `evalmc push` | Submit a Docker / ECR agent image |
+
+### ONNX (recommended)
+
+Graph is `obs → action`. No `policy.py` required.
+
+```bash
+evalmc submit-onnx model.onnx -t chasetag -p 1
+```
+
+ChaseTag tensors: input `obs` `[batch, 546]`, output `action` `[batch, 354]`.
+
+### Title-cycle nomination
+
+Before freeze, choose which scored in-window submission plays in finals.
+Without a nomination the platform uses your best score.
+
+```bash
+evalmc nominate -t chasetag -p 1 --submission-id 123
+```
+
+### Docker
+
+For custom Python around the policy. Ship `policy.py` and/or `model.onnx`
+under `/policy` (or `/app`):
+
+```python
+class Policy:
+    def reset(self):
+        ...
+    def get_action(self, obs):
+        ...  # 1-D float array, length 354 for ChaseTag
+```
+
+```bash
+evalmc push my-agent:tag -t chasetag -p 1
+```
+
+---
+
+ChaseTag 2026 is invitation-only. Registration and submission are rejected until
+the team is on the whitelist. Leaderboards stay public.

{evalmc-0.0.2 → evalmc-0.0.4}/pyproject.toml

@@ -4,22 +4,25 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "evalmc"
-version = "0.0.2"
-description = "MyoChallenge submission CLI"
+version = "0.0.4"
+description = "MyoChallenge CLI — submit ONNX and Docker policies trained with MyoSuite 3"
 readme = "README.md"
 license = "Apache-2.0"
-requires-python = ">=3.9"
+requires-python = ">=3.10"
 authors = [
     { name = "MyoChallenge Team", email = "caggiano@gmail.com" },
 ]
 classifiers = [
     "Development Status :: 4 - Beta",
-    "Intended Audience :: Developers",
+    "Intended Audience :: Science/Research",
     "Operating System :: OS Independent",
     "Programming Language :: Python :: 3",
-    "Programming Language :: Python :: 3.9",
     "Programming Language :: Python :: 3.10",
     "Programming Language :: Python :: 3.11",
+    "Programming Language :: Python :: 3.12",
+    "Programming Language :: Python :: 3.13",
+    "Programming Language :: Python :: 3.14",
+    "Topic :: Scientific/Engineering",
 ]
 dependencies = [
     "click>=8.0.0",
@@ -29,7 +32,9 @@ dependencies = [
 ]
 
 [project.urls]
-Homepage = "https://github.com/MyoHub/mcEvalProd-CLI"
+Homepage = "https://mc26.myosuite.org"
+MyoSuite = "https://www.myosuite.org"
+API = "https://api.mc.myosuite.org"
 
 [project.scripts]
 evalmc = "evalmc.main:main"

evalmc-0.0.2/PKG-INFO

@@ -1,60 +0,0 @@
-Metadata-Version: 2.4
-Name: evalmc
-Version: 0.0.2
-Summary: MyoChallenge submission CLI
-Author-email: MyoChallenge Team <caggiano@gmail.com>
-License-Expression: Apache-2.0
-Project-URL: Homepage, https://github.com/MyoHub/mcEvalProd-CLI
-Classifier: Development Status :: 4 - Beta
-Classifier: Intended Audience :: Developers
-Classifier: Operating System :: OS Independent
-Classifier: Programming Language :: Python :: 3
-Classifier: Programming Language :: Python :: 3.9
-Classifier: Programming Language :: Python :: 3.10
-Classifier: Programming Language :: Python :: 3.11
-Requires-Python: >=3.9
-Description-Content-Type: text/markdown
-License-File: LICENSE
-Requires-Dist: click>=8.0.0
-Requires-Dist: docker>=6.0.0
-Requires-Dist: requests>=2.25.0
-Requires-Dist: boto3>=1.26.0
-Dynamic: license-file
-
-# evalmc — MyoChallenge submission CLI
-
-```bash
-pip install -U evalmc
-```
-
-Default API host is `https://api.mc.myosuite.org`.
-
-## Setup
-
-```bash
-evalmc host --set-host https://api.mc.myosuite.org
-evalmc login          # email + auth token from mc26.myosuite.org
-evalmc status
-```
-
-## Submit
-
-**ONNX** (recommended for ChaseTag):
-
-```bash
-evalmc submit-onnx model.onnx -t chasetag -p 1
-```
-
-**Nominate** a scored in-window submission for title-cycle finals:
-
-```bash
-evalmc nominate -t chasetag -p 1 --submission-id 123
-```
-
-**Docker / ECR** (custom Python policies):
-
-```bash
-evalmc push my-agent:tag -t chasetag -p 1
-```
-
-Credentials live in `~/.myochallenge/` (`host_url`, `auth_token.json`).

evalmc-0.0.2/README.md

@@ -1,37 +0,0 @@
-# evalmc — MyoChallenge submission CLI
-
-```bash
-pip install -U evalmc
-```
-
-Default API host is `https://api.mc.myosuite.org`.
-
-## Setup
-
-```bash
-evalmc host --set-host https://api.mc.myosuite.org
-evalmc login          # email + auth token from mc26.myosuite.org
-evalmc status
-```
-
-## Submit
-
-**ONNX** (recommended for ChaseTag):
-
-```bash
-evalmc submit-onnx model.onnx -t chasetag -p 1
-```
-
-**Nominate** a scored in-window submission for title-cycle finals:
-
-```bash
-evalmc nominate -t chasetag -p 1 --submission-id 123
-```
-
-**Docker / ECR** (custom Python policies):
-
-```bash
-evalmc push my-agent:tag -t chasetag -p 1
-```
-
-Credentials live in `~/.myochallenge/` (`host_url`, `auth_token.json`).

evalmc-0.0.2/evalmc.egg-info/PKG-INFO

@@ -1,60 +0,0 @@
-Metadata-Version: 2.4
-Name: evalmc
-Version: 0.0.2
-Summary: MyoChallenge submission CLI
-Author-email: MyoChallenge Team <caggiano@gmail.com>
-License-Expression: Apache-2.0
-Project-URL: Homepage, https://github.com/MyoHub/mcEvalProd-CLI
-Classifier: Development Status :: 4 - Beta
-Classifier: Intended Audience :: Developers
-Classifier: Operating System :: OS Independent
-Classifier: Programming Language :: Python :: 3
-Classifier: Programming Language :: Python :: 3.9
-Classifier: Programming Language :: Python :: 3.10
-Classifier: Programming Language :: Python :: 3.11
-Requires-Python: >=3.9
-Description-Content-Type: text/markdown
-License-File: LICENSE
-Requires-Dist: click>=8.0.0
-Requires-Dist: docker>=6.0.0
-Requires-Dist: requests>=2.25.0
-Requires-Dist: boto3>=1.26.0
-Dynamic: license-file
-
-# evalmc — MyoChallenge submission CLI
-
-```bash
-pip install -U evalmc
-```
-
-Default API host is `https://api.mc.myosuite.org`.
-
-## Setup
-
-```bash
-evalmc host --set-host https://api.mc.myosuite.org
-evalmc login          # email + auth token from mc26.myosuite.org
-evalmc status
-```
-
-## Submit
-
-**ONNX** (recommended for ChaseTag):
-
-```bash
-evalmc submit-onnx model.onnx -t chasetag -p 1
-```
-
-**Nominate** a scored in-window submission for title-cycle finals:
-
-```bash
-evalmc nominate -t chasetag -p 1 --submission-id 123
-```
-
-**Docker / ECR** (custom Python policies):
-
-```bash
-evalmc push my-agent:tag -t chasetag -p 1
-```
-
-Credentials live in `~/.myochallenge/` (`host_url`, `auth_token.json`).