Updates for supporting CVDP Agentic subset#1744
Draft
arti4nvj wants to merge 10 commits into
Draft
Conversation
Add an ApptainerProvider implementing the SandboxProvider protocol via the local apptainer CLI: persistent instance lifecycle, exec with user/fakeroot mapping, bind-mount file transfer, status, readiness probe, and teardown. Register it under the name "apptainer" and add unit tests plus a README. Signed-off-by: Arti Jain <artij@nvidia.com>
Signed-off-by: Arti Jain <artij@nvidia.com>
Signed-off-by: Arti Jain <artij@nvidia.com>
Signed-off-by: Arti Jain <artij@nvidia.com>
Parse Claude Code's authoritative num_turns from the stream-json result event and include it in the returned metadata. Signed-off-by: Arti Jain <artij@nvidia.com>
Add the CVDP code-generation environment built on the Apptainer sandbox provider: resources server with harness execution, non-agentic and agentic cvdp_agent harnesses, configs, tests, and example dataset. Signed-off-by: Arti Jain <artij@nvidia.com>
…cvdp_resources_server # Conflicts: # resources_servers/cvdp/README.md
|
|
||
| There are two ways to drive this resources server: | ||
|
|
||
| - **Non-agentic** (`cvdp_agent`, `responses_api_agents/cvdp_agent/app.py`, config `configs/cvdp_agent.yaml`): the model emits the RTL directly in its text response; the server parses it out and runs the harness. |
Contributor
There was a problem hiding this comment.
it's a bit confusing to describe this path as non-agentic, considering cvdp_agent is itself an agent that we are using in the first scenario
| There are two ways to drive this resources server: | ||
|
|
||
| - **Non-agentic** (`cvdp_agent`, `responses_api_agents/cvdp_agent/app.py`, config `configs/cvdp_agent.yaml`): the model emits the RTL directly in its text response; the server parses it out and runs the harness. | ||
| - **Agentic** (`cvdp_agent_agentic`, `responses_api_agents/cvdp_agent/agentic_app.py`, config `configs/cvdp_agent_agentic.yaml`): runs Claude Code **inside** the EDA sim container so it can edit files on disk and self-test with the in-container EDA tools, then reports the files it wrote back to the server as `rtl_files` for grading. See `[responses_api_agents/cvdp_agent/](../../responses_api_agents/cvdp_agent/)`. |
Contributor
There was a problem hiding this comment.
I'd recommend we think about harness as first-class composable unit, describe this as illustration using Claude Code but could swap in other harnesses as well.
Contributor
There was a problem hiding this comment.
what's the rationale for splitting the verifier into two files app.py and the naming behind harness.py?
Contributor
|
would suggest considering an approach like this to reuse all agent harnesses with 0 code rewriting https://github.com/NVIDIA-NeMo/Gym/blob/main/responses_api_agents/anyterminal_agent/app.py#L190 |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Refactor the CVDP resources server to verify RTL using the Apptainer Provider sandbox instead of the previous Docker harness, and add an agentic CVDP agent.
harness.pysoapp.pyholds only the verifier logic.For n=1, for the agentic non-commerical subset, seeing 35.87% pass rate (compared to 40% from the original cvdp infra). For the non-agentic non-commerical subset, seeing 41.72% (in line with original cvdp infra).