Add VideoChat-Flash (OpenGVLab) language model support by anilmartha · Pull Request #2147 · microsoft/onnxruntime-genai

anilmartha · 2026-05-08T12:38:56Z

Summary

Adds support for OpenGVLab/VideoChat-Flash-Qwen2_5-7B_InternVideo2-1B, a video-language model whose architecture is VideoChatFlashQwenForCausalLM. The language backbone is standard Qwen2.5-7B (flat config, standard weight keys, 2D RoPE, rope_theta=1e6, 28 layers / 28 heads / 4 KV heads / hidden=3584). Because it does not use MRoPE, the new builder inherits from QwenModel directly rather than Qwen25VLTextModel.

Changes

Python builder (`src/python/py/models/`)

builders/qwen.py — New VideoChatFlashQwenModel subclass of QwenModel.
- Overrides make_genai_config() to bypass the model's custom remote code (which imports av, cv2, decord, imageio even for config-only loads) by writing a temporary Qwen2Config-based config.json and patching genai_config.json["model"]["type"] back to videochat_flash_qwen.
- Overrides load_weights() to load the LM backbone directly via Qwen2ForCausalLM.from_pretrained(), avoiding video deps.
- Sets model_type = "videochat_flash_qwen" and hf_remote = False.
builders/__init__.py — Exports VideoChatFlashQwenModel.
builder.py —
- Detects VideoChatFlashQwenForCausalLM by peeking at raw config.json (local dir or via hf_hub_download) before calling AutoConfig, then loads as Qwen2Config to avoid pulling in video libraries.
- Routes VideoChatFlashQwenForCausalLM → VideoChatFlashQwenModel and defaults exclude_embeds=True (the text decoder receives inputs_embeds from the embedding merger). exclude_embeds is now opt-in (guarded with not in extra_options) so standalone input_ids export still works.

C++ runtime (`src/`)

models/videochat_flash_processor.{h,cpp} (new) — VideoChatFlashProcessor runs ORT-Extensions image preprocessing, transposes HWC→CHW, reshapes to [1, num_frames, C, H, W], casts pixel values to the session input dtype, builds input_ids by replacing each <|vision_start|>…<|vision_end|> block with a fixed number of <|image_pad|> tokens, and emits image_grid_thw so GetImageFeatureBatchSize can determine num_images after the pixel_values name remap.
models/model.cpp — Registers "videochat_flash_qwen" → VideoChatFlashProcessor in MultiModalProcessor.
models/model_type.h — Adds "videochat_flash_qwen" to the IsVLM() list (size 6 → 7).
config.{h,cpp} — Adds Vision::num_visual_tokens (default 0) — fixed visual tokens per image; 0 falls back to computing from image_grid_thw. Required (>0) by the new processor.

Files changed

File	Change
`src/config.cpp`	Parse `num_visual_tokens` from vision config
`src/config.h`	Add `num_visual_tokens` field on `Vision`
`src/models/model.cpp`	Register `VideoChatFlashProcessor`
`src/models/model.h`	Header / license update
`src/models/model_type.h`	Add `videochat_flash_qwen` to VLM list
`src/models/videochat_flash_processor.cpp`	New — image processor
`src/models/videochat_flash_processor.h`	New — header
`src/python/py/models/builder.py`	Detect VCF arch, bypass remote-code video deps, route to new builder
`src/python/py/models/builders/__init__.py`	Export `VideoChatFlashQwenModel`
`src/python/py/models/builders/qwen.py`	Add `VideoChatFlashQwenModel`

Validation

Text-only export of the Qwen2.5-7B backbone via python -m onnxruntime_genai.models.builder -m OpenGVLab/VideoChat-Flash-Qwen2_5-7B_InternVideo2-1B -o <dir> -p fp32 -e cpu --extra_options exclude_embeds=false succeeds without av/cv2/decord/imageio installed.
Standalone inference on the exported vcf-oga-fp32-standalone model produces correct answers across 4 batched QA prompts (e.g. Paris, 56, ONNX Runtime description) using generator.append_tokens() on the Qwen2.5-7B backbone.

Test plan

Build OGA with the new processor compiled in and confirm IsVLM("videochat_flash_qwen") returns true.
Export the text decoder with exclude_embeds=false and run text-only inference end-to-end.
Confirm genai_config.json contains "type": "videochat_flash_qwen" (full VLM) or "type": "qwen2" (text-only standalone).
Image-input smoke test once vision.onnx and embedding.onnx land.

Adds text decoder export support for OpenGVLab/VideoChat-Flash-Qwen2_5-7B_InternVideo2-1B. Architecture: VideoChatFlashQwenForCausalLM is a VLM whose language backbone is standard Qwen2.5-7B (flat config, standard weight keys, 2D RoPE, rope_theta=1e6, 28L / 28h / 4kv / hidden=3584). The model does NOT use MRoPE, so the builder inherits QwenModel directly rather than Qwen25VLTextModel. Changes: - src/python/py/models/builders/qwen.py: Add VideoChatFlashQwenModel subclass of QwenModel. Sets exclude_embeds=True (text decoder receives inputs_embeds from the embedding merger) and model_type="videochat_flash_qwen". - src/python/py/models/builders/__init__.py: Export VideoChatFlashQwenModel. - src/python/py/models/builder.py: Map "VideoChatFlashQwenForCausalLM" architecture string to VideoChatFlashQwenModel with exclude_embeds=True. - src/models/model_type.h: Register "videochat_flash_qwen" in IsVLM() (size 6->7). - examples/python/videochat-flash/builder.py: New example export script. Phase 1 (this PR): text decoder only. Vision encoder (InternVideo2-1B) and embedding merger export are Phase 2 (scaffolded as TODOs). Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

…inference support - builder.py (example): use local OGA builder via sys.path; fix create_model args when no local input dir; move prepare_model() inside Phase-2 block (trust_remote_code=False is safe for config-only loads) - builder.py (core): add VCF config bypass using hf_hub_download + Qwen2Config to avoid av/cv2/decord/imageio imports triggered by AutoConfig; set config._name_or_path so load_weights() resolves the model correctly; make exclude_embeds opt-in (guard with 'not in extra_options') so standalone (input_ids) export also works - qwen.py: add make_genai_config() override that writes a temp Qwen2Config and patches genai_config.json type back to videochat_flash_qwen; add load_weights() override using Qwen2ForCausalLM.from_pretrained() directly Validated: text-only inference with vcf-oga-fp32-standalone/ produces correct answers (Paris, 56, ONNX Runtime description) using append_tokens() API on the exported Qwen2.5-7B backbone. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

examples/python/videochat-flash/run.py: - Text-only inference for the exported OGA model - Uses HF AutoTokenizer (og.Tokenizer fails for TokenizersBackend models) - Feeds tokens via generator.append_tokens(np.int32 array) - --batch flag runs 4 built-in QA prompts for quick validation - --prompt / --max-length for custom single-prompt runs - Notes in docstring on genai_config.json type patch workaround for installed OGA binaries that predate videochat_flash_qwen support Validated: all 4 batch prompts produce correct answers on vcf-oga-fp32-standalone. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

…xt-only mode Root cause of teammate's inference failure: 1. builder.py set type=videochat_flash_qwen even in --text-only mode. OGA loads this via MultiModalLanguageModel which requires vision.onnx + embedding.onnx (not present in text-only export) → 'File doesn't exist' error on og.Model(). 2. exclude_embeds defaulted to true → decoder expected inputs_embeds, but run.py feeds token IDs via append_tokens() → 'input not found' error. Fix: - builder.py export_text_model(): pass exclude_embeds=false in text-only mode (decoder uses input_ids for standalone inference). - builder.py update_genai_config(): set type=qwen2 in text-only mode so OGA loads the model as a plain decoder (LM backbone is identical to Qwen2.5-7B). type=videochat_flash_qwen is reserved for Phase 2 full VLM pipeline. - run.py: simplify docstring now that no manual patching is needed. Validated with compiled OGA build (onnxruntime-genai conda env, Python 3.11): all 4 batch prompts produce correct answers without any manual config changes. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

Documents Phase 1 (text decoder, done) and Phase 2 (vision + embedding, TODO): - Architecture overview: InternVideo2-1B ViT → mm_projector → embedding merger → Qwen2.5-7B - Phase 1 summary: what was implemented, design decisions, usage instructions - Phase 2 roadmap: vision.onnx export (39-block ViT + MLP projector), embedding.onnx (embed_tokens + image-pad replacement), genai_config.json wiring, video preprocessing, and optional in-LLM HiCo compression (llm_compress_layer_list) - Key challenges: 3D spatiotemporal attention ONNX compat, ToMe token merging ops, dynamic temporal position embeddings - File map and reference links Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

…ipt+ config

VishalX · 2026-05-08T18:10:13Z

Olive Recipe: microsoft/olive-recipes#406

anilmartha · 2026-05-11T09:52:27Z

@microsoft-github-policy-service agree company="AMD"

Copilot

Pull request overview

Adds first-class support for the OpenGVLab VideoChat-Flash Qwen-based video-language model by extending the Python model builder to export its (text) decoder without pulling video dependencies, and extending the C++ runtime to recognize/run the new multimodal model type.

Changes:

Python: detect VideoChatFlashQwenForCausalLM early from raw config.json, route to a new VideoChatFlashQwenModel, and avoid trust_remote_code video-library imports.
C++: introduce VideoChatFlashProcessor and register videochat_flash_qwen as a VLM model type/processor.
Config: add and parse vision.num_visual_tokens used by the new processor for fixed visual-token padding.

Reviewed changes

Copilot reviewed 10 out of 10 changed files in this pull request and generated 4 comments.

Show a summary per file

File	Description
src/python/py/models/builders/qwen.py	Adds `VideoChatFlashQwenModel` subclass to export the LM backbone while bypassing remote-code video deps.
src/python/py/models/builders/init.py	Exports the new builder class from the builders package.
src/python/py/models/builder.py	Adds config “peek” detection for VideoChat-Flash and routes to the new builder with default `exclude_embeds=True`.
src/models/videochat_flash_processor.h	Declares `VideoChatFlashProcessor` (new multimodal processor).
src/models/videochat_flash_processor.cpp	Implements preprocessing/token padding + emits `image_grid_thw` for batching inference.
src/models/model.h	Updates header/license comments.
src/models/model.cpp	Registers `videochat_flash_qwen` → `VideoChatFlashProcessor` in the multimodal processor factory.
src/models/model_type.h	Adds `videochat_flash_qwen` to the VLM model-type list.
src/config.h	Adds `vision.num_visual_tokens` to config schema.
src/config.cpp	Parses `vision.num_visual_tokens` from `genai_config.json`.

Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com>

- builder.py: introduce _has_vcf_architecture() helper that handles architectures as a list, a bare string, or absent/unknown type, then use it in both the local-dir and HF-hub code paths (Copilot comment #2) - config.h: correct misleading comment on num_visual_tokens — remove the false "0 = compute from image_grid_thw" claim; the field must be > 0 for videochat_flash_qwen (Copilot comment #3) - videochat_flash_processor.cpp: add clarifying comment to empty catch block — exception is expected when only the language decoder session is loaded (Copilot comment microsoft#4) - builders/__init__.py: move VideoChatFlashQwenModel to its correct alphabetical position in __all__ (after SmolLM3Model, before WhisperModel) per kunal-vaishnavi comment microsoft#6 Co-Authored-By: Claude Sonnet 4 <noreply@anthropic.com>

anilmartha · 2026-05-18T10:25:54Z

Hi @kunal-vaishnavi
Could you please review it again?

Copilot

Copilot encountered an error and was unable to review this pull request. You can try again by re-requesting a review.

Copilot

Copilot encountered an error and was unable to review this pull request. You can try again by re-requesting a review.

kunal-vaishnavi · 2026-05-19T22:41:54Z

+  ort_extensions::OrtxObjectPtr<OrtxTensorResult> result;
+  CheckResult(OrtxImagePreProcess(processor_.get(), images->images_.get(), result.ToBeAssigned()));
+
+  OrtxTensor* pixel_values = nullptr;


This appears to be a memory leak. Can you fix this? There was a recent PR that fixes this issue.

kunal-vaishnavi · 2026-05-19T22:51:49Z

+                             std::to_string(pv_ndims) + " (expected 3 or 4)");
+  }
+
+  // Transpose HWC → CHW and reshape to [1, num_frames, C, H, W]


Let's do this transpose as another step inside ORT Extensions before returning the tensor to ORT GenAI.

kunal-vaishnavi · 2026-05-19T22:54:42Z

+    }
+  }
+
+  auto converted_pv = ConvertPixelValues(*float_tensor, pixel_values_type_, allocator);


Rather than creating a new method called ConvertPixelValues, can we re-use existing APIs that already convert dtypes? Here is an example.

onnxruntime-genai/src/models/whisper_processor.cpp

Lines 33 to 39 in ca9ecd2

if (audio_features_type_ == ONNX_TENSOR_ELEMENT_DATA_TYPE_FLOAT) {

named_tensors->emplace(std::string(Config::Defaults::AudioFeaturesName),

std::make_shared<Tensor>(ProcessTensor<float>(mel.get(), allocator)));

} else {

named_tensors->emplace(std::string(Config::Defaults::AudioFeaturesName),

std::make_shared<Tensor>(ProcessTensor<Ort::Float16_t>(mel.get(), allocator)));

}

kunal-vaishnavi · 2026-05-19T22:57:52Z

+    grid_ptr[i * 3 + 1] = height;
+    grid_ptr[i * 3 + 2] = width;
+  }
+  named_tensors->emplace("image_grid_thw",


If there are new input names, let's add their default names in the config (e.g. Config::Defaults::ImageGridThw) and reference them here.

kunal-vaishnavi · 2026-05-19T23:01:33Z

    Qwen25VLTextModel,
    Qwen35TextModel,
    QwenModel,
+    VideoChatFlashQwenModel,


Let's preserve any existing alphabetical orders when inserting new items.

kunal-vaishnavi · 2026-05-19T23:35:12Z

+        # hf_remote=False so base class helpers (make_genai_config,
+        # save_processing) use standard, non-remote-code paths.
+        extra_options = dict(extra_options) if extra_options else {}
+        extra_options["hf_remote"] = False


To be consistent with other models, let's set any default extra options flags during architecture parsing. For example:

onnxruntime-genai/src/python/py/models/builder.py

Lines 292 to 299 in ca9ecd2

elif config.architectures[0] == "Qwen2_5_VLForConditionalGeneration":

text_config = config.text_config

for key in text_config:

if not hasattr(config, key):

setattr(config, key, getattr(text_config, key))

print("WARNING: This is only generating the text component of the model. Setting `--extra_options exclude_embeds=true` by default.")

extra_options["exclude_embeds"] = True

onnx_model = Qwen25VLTextModel(config, io_dtype, onnx_dtype, execution_provider, cache_dir, extra_options)

kunal-vaishnavi · 2026-05-19T23:39:09Z

+        # establishes these attributes on first assignment instead of us
+        # overwriting inherited values.
+        #
+        # The custom remote code requires video libraries (av, cv2, decord,


It's fine if video libraries need to be temporarily installed. These libraries are on a per-model basis and are based on Hugging Face's requirements for loading a model. For example, Phi-4 mm requires PEFT for loading the model and we don't remove those library dependencies even though they are not technically needed for producing the ONNX model.

kunal-vaishnavi · 2026-05-19T23:40:14Z

+
+        super().__init__(config, io_dtype, onnx_dtype, ep, cache_dir, extra_options)
+
+    def make_genai_config(self, model_name_or_path, extra_kwargs, out_dir):


Same comment as here. We should not need to override this method at all.

kunal-vaishnavi · 2026-05-19T23:42:16Z

+
+        from transformers import Qwen2Config
+
+        vcf_config = Qwen2Config.from_pretrained(model_name_or_path, token=self.hf_token, **extra_kwargs)


If you want to avoid loading the video library imports by using a Qwen2Config for importing, can you create a generic load_config method in the base class (similar to load_weights below)? Then, we can call that method in the various locations that require a config. That method can then be overridden in this class to be Qwen-2 specific.

kunal-vaishnavi · 2026-05-19T23:43:47Z

    hf_remote = extra_options.get("hf_remote", True)

-    config = AutoConfig.from_pretrained(hf_name, token=hf_token, trust_remote_code=hf_remote, **extra_kwargs)
+    # VideoChat-Flash uses custom remote code that imports heavy video libraries


All of the below changes should be reverted. They are too specific for this model and this section needs to remain generic. We can try the load_config suggestion from earlier to resolve this.

kunal-vaishnavi · 2026-05-19T23:49:58Z

        onnx_model = Phi4MMModel(config, io_dtype, onnx_dtype, execution_provider, cache_dir, extra_options)
    elif config.architectures[0] == "Qwen2ForCausalLM":
        onnx_model = QwenModel(config, io_dtype, onnx_dtype, execution_provider, cache_dir, extra_options)
+    elif config.architectures[0] == "VideoChatFlashQwenForCausalLM":


Can we remove the if condition?

onnxruntime-genai/src/python/py/models/builder.py

Lines 292 to 299 in f32444d

elif config.architectures[0] == "Qwen2_5_VLForConditionalGeneration":

text_config = config.text_config

for key in text_config:

if not hasattr(config, key):

setattr(config, key, getattr(text_config, key))

print("WARNING: This is only generating the text component of the model. Setting `--extra_options exclude_embeds=true` by default.")

extra_options["exclude_embeds"] = True

onnx_model = Qwen25VLTextModel(config, io_dtype, onnx_dtype, execution_provider, cache_dir, extra_options)

Multi-modal models do not need to check this since the embedding ONNX model will go from input_ids to inputs_embeds.

…ated

Address PR microsoft#2147 review: keep the config loading section generic by removing the VideoChat-Flash architecture peek/Qwen2Config workaround. Use the standard AutoConfig.from_pretrained path for all models. Also remove the forced hf_remote=False in the VCF dispatch branch so HF model dependencies (video libraries) are loaded normally, consistent with how other models handle their library requirements. Co-authored-by: Cursor <cursoragent@cursor.com>

anilmartha · 2026-05-21T10:01:38Z

Hi @kunal-vaishnavi
I have addressed the review comments. Could you please take another look?

+        # and genai_config.json. Same pattern as Qwen3VLTextModel.
+        # Base class transforms this to "videochat_flash_qwen" via:
+        #   model_type[:model_type.find("For")].lower()
+        self.model_type = "VideoChat_Flash_QwenForCausalLM"


amdrajeevp1 and others added 16 commits March 24, 2026 17:53

Fix text only model

2735c49

Add preprocessor pipeline with adjusting qwenimageprocessor + add scr…

799c6f3

…ipt+ config

Add full standalone ort pipeline

036ea67

Add videochat_flash_qwen image processor

21b5a46

Remove example folder

36ff417

Revert the changes

5dc8084

Add license

f7e7b5d

Add spdx-mit

4493383

Add spdx-mit

fcf7278

Merge branch 'upstream_main' into add-opengv-support1

99893f6

Merge branch 'main' into add-opengv-support_updated

4d12135

anilmartha marked this pull request as ready for review May 11, 2026 09:51

anilmartha requested a review from a team as a code owner May 11, 2026 09:51

Copilot AI review requested due to automatic review settings May 11, 2026 09:51

Copilot started reviewing on behalf of anilmartha May 11, 2026 09:52 View session

Copilot AI reviewed May 11, 2026

View reviewed changes

Comment thread src/python/py/models/builders/__init__.py Outdated

Comment thread src/python/py/models/builder.py Outdated

Comment thread src/config.h Outdated

Comment thread src/models/videochat_flash_processor.cpp

xieofxie mentioned this pull request May 12, 2026

Add VideoChat-Flash-Qwen2_5-7B_InternVideo2-1B recipe microsoft/olive-recipes#406

Open

kunal-vaishnavi reviewed May 15, 2026

View reviewed changes

Comment thread src/python/py/models/builders/__init__.py Outdated

anilmartha and others added 2 commits May 18, 2026 11:03

Apply review changes

47ebb9e

Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com>

github-advanced-security AI found potential problems May 18, 2026

View reviewed changes

Comment thread src/python/py/models/builders/qwen.py Fixed

Comment thread src/python/py/models/builders/qwen.py Fixed

Comment thread src/python/py/models/builder.py Fixed

Address CodeQL findings

9315ab7

anilmartha requested a review from Copilot May 18, 2026 14:42

Copilot started reviewing on behalf of anilmartha May 18, 2026 14:45 View session

Copilot AI reviewed May 18, 2026

View reviewed changes

anilmartha requested a review from Copilot May 18, 2026 14:53

Copilot started reviewing on behalf of anilmartha May 18, 2026 14:58 View session

Copilot AI reviewed May 18, 2026

View reviewed changes

kunal-vaishnavi added the 0.14.0 label May 18, 2026

Fix linter

4faa201

kunal-vaishnavi reviewed May 19, 2026

View reviewed changes

Comment thread src/python/py/models/builders/qwen.py Outdated

kunal-vaishnavi reviewed May 19, 2026

View reviewed changes

Anil Kumar Martha and others added 4 commits May 20, 2026 11:48

Merge remote-tracking branch 'anilm/main' into add-opengv-support_upd…

38fdd6b

…ated

Address review comments

23d25ea

Fix linter

cb226b7

github-advanced-security AI found potential problems May 21, 2026

View reviewed changes

Comment thread src/python/py/models/builders/qwen.py

# and genai_config.json. Same pattern as Qwen3VLTextModel.

# Base class transforms this to "videochat_flash_qwen" via:

# model_type[:model_type.find("For")].lower()

self.model_type = "VideoChat_Flash_QwenForCausalLM"

	if (audio_features_type_ == ONNX_TENSOR_ELEMENT_DATA_TYPE_FLOAT) {
	named_tensors->emplace(std::string(Config::Defaults::AudioFeaturesName),
	std::make_shared<Tensor>(ProcessTensor<float>(mel.get(), allocator)));
	} else {
	named_tensors->emplace(std::string(Config::Defaults::AudioFeaturesName),
	std::make_shared<Tensor>(ProcessTensor<Ort::Float16_t>(mel.get(), allocator)));
	}

	elif config.architectures[0] == "Qwen2_5_VLForConditionalGeneration":
	text_config = config.text_config
	for key in text_config:
	if not hasattr(config, key):
	setattr(config, key, getattr(text_config, key))
	print("WARNING: This is only generating the text component of the model. Setting `--extra_options exclude_embeds=true` by default.")
	extra_options["exclude_embeds"] = True
	onnx_model = Qwen25VLTextModel(config, io_dtype, onnx_dtype, execution_provider, cache_dir, extra_options)


		super().__init__(config, io_dtype, onnx_dtype, ep, cache_dir, extra_options)

		def make_genai_config(self, model_name_or_path, extra_kwargs, out_dir):


		from transformers import Qwen2Config

		vcf_config = Qwen2Config.from_pretrained(model_name_or_path, token=self.hf_token, **extra_kwargs)

Conversation

anilmartha commented May 8, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Summary

Changes

Python builder (src/python/py/models/)

C++ runtime (src/)

Files changed

Validation

Test plan

Uh oh!

VishalX commented May 8, 2026

Uh oh!

anilmartha commented May 11, 2026

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Reviewed changes

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

anilmartha commented May 18, 2026

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

anilmartha commented May 21, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

6 participants

anilmartha commented May 8, 2026 •

edited

Loading

Python builder (`src/python/py/models/`)

C++ runtime (`src/`)