Skip to content

Migrate decoder configs to generic runtime type - #624

Merged
justinchuby merged 1 commit into
mainfrom
justinchuby-migrate-generic-decoder-configs
Aug 25, 2026
Merged

Migrate decoder configs to generic runtime type#624
justinchuby merged 1 commit into
mainfrom
justinchuby-migrate-generic-decoder-configs

Conversation

@justinchuby

Copy link
Copy Markdown
Member

Summary

  • normalize every graph-representable decoder-only ORT GenAI config emitted by Mobius to model.type: "decoder"
  • preserve specialized types only for GPT-2's unrepresentable cache ABI, LFM2's convolution cache, Phi-3-family LongRoPE, multimodal/encoder-decoder pipelines, and auxiliary sidecar topologies
  • update the Gemma4 decoder-only example, CLI docs, and ORT GenAI config skill; keep Add ORT GenAI end-to-end CI #618's OGA 0.15.2 CI and evidence unchanged

Validation

  • 235 passed focused ORT GenAI config tests
  • 7717 passed, 56 skipped, 1 subtest passed broad non-integration suite
  • 37 passed GGUF runtime tests
  • 2 passed network-free OGA 0.15.2 fast E2E
  • 1 passed pinned real SmolLM F16 CPU OGA 0.15.2 E2E
  • Sphinx documentation build
  • full lintrunner formatting/lint
  • two GPT-5.6 Sol medium reviews; final review reported no findings

Base: e591ecfaae00d312cb64c4334b61aff0a9e87f05

Normalize every graph-representable decoder-only ORT GenAI config to the architecture-neutral decoder type. Preserve only runtime-specific cache behavior, LongRoPE, and auxiliary pipeline topologies, and keep emitted compatibility metadata aligned with the generated config.\n\nUpdate focused coverage, the Gemma4 text example, CLI documentation, and the ORT GenAI config skill to make the generic decoder contract the default.\n\nCo-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com>

Signed-off-by: Justin Chu <justinchuby@users.noreply.github.com>
@justinchuby
justinchuby requested review from a team and a lite review from Copilot August 25, 2026 19:52
@justinchuby
justinchuby merged commit 04b7e3f into main Aug 25, 2026
15 of 24 checks passed
@justinchuby
justinchuby deleted the justinchuby-migrate-generic-decoder-configs branch August 25, 2026 19:52
@github-actions

Copy link
Copy Markdown

🏗️ Architecture Diff

Comparing e591ecfb030c36

Model Sub-model Changes Status

No architecture changes detected.


Legend: ⚪ No change · 🔵 Minor (attrs/inits) · 🟡 Moderate (nodes added/removed) · 🔴 Major (interface changed)

@github-actions

Copy link
Copy Markdown

Performance Comparison

Comparing e591ecfb030c36

Model Metric Baseline Current Delta
bert (feature-extraction) model_size_bytes 359 KB 359 KB +0.0%
bert (feature-extraction) num_nodes 68 68 +0.0%
falcon model_size_bytes 364 KB 364 KB +0.0%
falcon num_nodes 66 66 +0.0%
gemma2 model_size_bytes 428 KB 428 KB +0.0%
gemma2 num_nodes 105 105 +0.0%
gpt2 model_size_bytes 388 KB 388 KB +0.0%
gpt2 num_nodes 54 54 +0.0%
llama model_size_bytes 425 KB 425 KB +0.0%
llama num_nodes 60 60 +0.0%
llama (static-cache) model_size_bytes 425 KB 425 KB +0.0%
llama (static-cache) num_nodes 56 56 +0.0%
mamba (ssm-text-generation) model_size_bytes 296 KB 296 KB +0.0%
mamba (ssm-text-generation) num_nodes 94 94 +0.0%
phi3 model_size_bytes 421 KB 421 KB +0.0%
phi3 num_nodes 58 58 +0.0%
phi3 (static-cache) model_size_bytes 421 KB 421 KB +0.0%
phi3 (static-cache) num_nodes 54 54 +0.0%
qwen2 model_size_bytes 425 KB 425 KB +0.0%
qwen2 num_nodes 60 60 +0.0%
qwen2 (static-cache) model_size_bytes 425 KB 425 KB +0.0%
qwen2 (static-cache) num_nodes 56 56 +0.0%
qwen3_5_moe (hybrid-text-generation) model_size_bytes 506 KB 506 KB +0.0%
qwen3_5_moe (hybrid-text-generation) num_nodes 265 265 +0.0%
qwen3_5_text (hybrid-text-generation) model_size_bytes 458 KB 458 KB +0.0%
qwen3_5_text (hybrid-text-generation) num_nodes 127 127 +0.0%
qwen3_5_vl (hybrid-qwen-vl) model_size_bytes 977 KB 977 KB +0.0%
qwen3_5_vl (hybrid-qwen-vl) num_nodes 429 429 +0.0%
t5 (seq2seq) model_size_bytes 836 KB 836 KB +0.0%
t5 (seq2seq) num_nodes 176 176 +0.0%
whisper (speech-to-text) model_size_bytes 1008 KB 1008 KB +0.0%
whisper (speech-to-text) num_nodes 128 128 +0.0%

No performance regressions.

Copilot AI left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Pull request overview

This PR standardizes Mobius-emitted ORT GenAI decoder-only configs to use the generic runtime model.type: "decoder" whenever the exported graph ABI is representable by the generic decoder contract, while preserving specialized runtime types only for known ABI/topology exceptions (e.g., GPT-2 cache ABI, LFM2 conv cache, Phi-3 LongRoPE, multimodal/pipeline/sidecar packages).

Changes:

  • Normalize decoder-only genai_config.json emission to model.type = "decoder", with explicit exception handling (specialized ABIs, LongRoPE, and specialized topologies).
  • Update auto-export model type selection to preserve Phi-3-family types only when rope_type == "longrope", and propagate “specialized topology” into config generation.
  • Refresh tests, docs, and examples to reflect the new generic decoder type behavior.

Reviewed changes

Copilot reviewed 9 out of 9 changed files in this pull request and generated 1 comment.

Show a summary per file
File Description
src/mobius/integrations/ort_genai/genai_config.py Adds normalization logic for emitted model.type with explicit exception gates (multimodal/topology, specialized ABIs, LongRoPE).
src/mobius/integrations/ort_genai/genai_config_test.py Updates expectations and adds coverage for normalization vs preserved specialized types/topologies.
src/mobius/integrations/ort_genai/auto_export.py Adjusts type selection (Phi-3 LongRoPE gating), threads topology flag into config generation, and aligns runtime compatibility metadata with emitted type.
src/mobius/integrations/ort_genai/auto_export_test.py Adds/updates tests for Phi-3 LongRoPE gating and sidecar topology type preservation.
examples/gemma4/ort_genai/text/genai_config.json Updates Gemma4 text-only example to emit model.type: "decoder".
examples/gemma4/ort_genai/README.md Updates narrative/table to reflect generic decoder for text-only and specialized type retention for multimodal.
docs/cli_reference.md Updates CLI docs describing when decoder is emitted and when specialized types remain.
.agents/skills/ort-genai-config/SKILL.md Updates the ORT GenAI config skill to reflect the new normalization rule and exception list.
.agents/skills/ort-genai-config/references/genai-config-fields.md Updates field documentation for model.type semantics (generic decoder vs explicit exceptions).

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.

Comment on lines 2031 to 2037
result["genai_config"] = genai_path
with open(genai_path, encoding="utf-8") as handle:
emitted_model_type = json.load(handle)["model"]["type"]
compatibility_path = _write_runtime_compatibility(
directory,
model_type=ort_model_type,
model_type=emitted_model_type,
runtime_version=runtime_version,
justinchuby added a commit that referenced this pull request Aug 25, 2026
## Summary

- add a dedicated Kimi-K3 config, model, and heterogeneous-state task
with exact KDA/NoPE gated-MLA scheduling, AttnRes mixing, SiTU latent
MoE, shared experts, and an untied output head
- import pinned llama.cpp `kimi-k3` GGUF metadata and tensors with
strict metadata/tensor/shape/storage closure and malformed-input
rejection
- preserve lossless quantization routes under #609: separate rank-3 MLA
projections fail closed unless explicitly dequantized, while fused Q4_0
KV-B is split by exact packed-row reordering across weights, scales, and
zero-points
- preserve valid KDA convolution history across padding, validate
kernel/state/task contracts, and cover replay, reorder, float/quantized
import, roundtrip, CLI, and negative paths
- keep generic OGA runtime packaging truthfully deferred under #605
because released cache schemas cannot represent the heterogeneous state
ABI

## Reconstruction

Reconstructed after #619 was admin squash-merged. The PR contains only
the four-commit Kimi-K3 delta and review follow-ups on live-main base
`04b7e3f6f2d9de5d5741aafb8fa6375a18eee693`. This preserves #611
graph-driven ORT GenAI decoder configs, #618 ORT GenAI end-to-end CI,
#624 generic decoder config migration, and the Kimi Linear, MiniMax,
runtime, and fail-closed quantization changes already on main. `git
range-diff` reports all four replayed commits as patch-identical (`=`).

## Validation

- Kimi-K3 model and GGUF tests: 30 passed
- focused builder/build-graph Kimi-K3 checks: 4 passed, 1900 deselected
- affected generic ORT config/runtime/E2E tests: 250 passed
- affected OGA metadata tests: 94 passed, 1 skipped
- broad non-integration suite: 7765 passed, 56 skipped, 1 subtest passed
- repository lintrunner passed
- final GPT-5.6 Sol medium review reported no findings

## Real-checkpoint note

The pinned
`yujiepan/kimi-k3-tiny-random@a5c86ee03f07f7b141508b0108304a1447fbb345`
checkpoint was assessed, but macOS cannot run its required `fla-core`
kernels and its selective compressed-tensors MXFP4 expert representation
is unsupported by the generic HF loader. That format fails explicitly
rather than producing an incorrectly quantized graph.

---------

Signed-off-by: Justin Chu <justinchuby@users.noreply.github.com>
Co-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com>
justinchuby added a commit that referenced this pull request Aug 25, 2026
## Summary

- clean up still-valid typing, documentation, test robustness,
error-message,
  maintainability, and performance findings left by the GGUF PR stack
- preserve the behavioral fixes merged in #625-#632
- hash reused multi-GB GGUF sources once, at the final pre-publication
integrity
  gate, while retaining cheap identity checks around staging

Exact base: `2db9d33debdc254d879a51b14434c9a81c230f4f`

Exact head: `4541ed2bc9d2ab4227484510b4a85b2d9113eb25`

## Reconstructed original 17-item low-priority tranche

The persisted audit retained only the totals, so this list was
reconstructed
from the live threads and current source. All but the already-fixed #596
comment are addressed in this PR.

| PR | Comment | Disposition |
|---|---:|---|
| #550 | 3837144656 | Implemented: correct tuple return annotation |
| #552 | 3837389183 | Implemented: multichannel waveform shape docs |
| #559 | 3837716261 | Implemented: name-based cache assertions |
| #573 | 3854308350 | Implemented: one final GGUF hash, with integrity
regression coverage |
| #574 | 3854345521 | Implemented: `TensorRole \| None` typing |
| #574 | 3854345597 | Implemented: removed obsolete verdict filtering |
| #577 | 3854472859 | Implemented: documented SSM sequence length |
| #578 | 3843112613 | Implemented: documented F64 passthrough |
| #579 | 3854518332 | Implemented: generalized fused-projection error |
| #580 | 3854576300 | Implemented: removed brittle node counts |
| #583 | 3854681386 | Implemented: documented conditional draft outputs
|
| #587 | 3854816872 | Implemented: corrected MTP output contract docs |
| #596 | 3846110059 | Already fixed on base: unambiguous GQA bias
comment |
| #600 | 3855174281 | Implemented: metadata-count-only MTP error |
| #607 | 3855776048 | Implemented: stable route-field assertions |
| #607 | 3855776086 | Implemented: public tensor iterator |
| #609 | 3856486136 | Implemented: fail-closed LM-head comment |

## Current unresolved-thread disposition

This covers all 48 Copilot threads returned by the reproducible
#600-#630
query. The one human #623 thread is excluded.

| PR | Comment | Current-main disposition and evidence |
|---|---:|---|
| #600 | 3855174177 | Already fixed by #629: package cycle and
reserved-sidecar validation |
| #600 | 3855174238 | Already fixed by #629: explicit MTP sidecar
naming/loading |
| #600 | 3855174281 | Implemented here: error no longer invents an
observed block count |
| #602 | 3848155961 | Outside exact stack; already fixed: top-level
`expert_dtype` is classified before early return |
| #602 | 3848155983 | Outside exact stack; still-valid behavioral
block-quant validation, unchanged |
| #602 | 3848156000 | Outside exact stack; still-valid truncated-read
behavioral finding, unchanged |
| #602 | 3848156022 | Outside exact stack; still-valid descriptor
byte/dtype validation, unchanged |
| #602 | 3848156040 | Outside exact stack; still-valid expert-bank
payload validation, unchanged |
| #603 | 3855249765 | Already fixed by #630: runtime preflight preserves
shard sets |
| #603 | 3855249840 | Already fixed by #630: success output follows
durable runtime publication |
| #604 | 3855343082 | Implemented here: graph-only MTP persistence
distinguished from runtime rejection |
| #604 | 3855343131 | Already fixed by #630: runtime success messages
are atomic |
| #607 | 3855776001 | Implemented here: missing generation golden skips
before provenance read |
| #607 | 3855776048 | Implemented here: only stable route fields are
asserted |
| #607 | 3855776086 | Implemented here: tensor count uses
`tensor_items_raw()` |
| #608 | 3856079371 | Still-valid behavioral cache-symlink containment
finding; unchanged |
| #608 | 3856079415 | Still-valid behavioral lowercase-digest validation
finding; unchanged |
| #609 | 3856486136 | Implemented here: comment matches value-preserving
policy |
| #610 | 3855541683 | Already fixed by #628: Falcon bias precedence is
explicit |
| #610 | 3855541761 | Already fixed by #628: CTRL tiny config exercises
projection biases |
| #611 | 3856595840 | Implemented here: runtime test resolves the
distribution providing the module |
| #612 | 3855677383 | Already fixed by #625: supported-version
endianness detection |
| #612 | 3855677427 | Implemented here: shared `INT64_MAX` sentinel |
| #612 | 3855677460 | Implemented here: shared PLaMo2 width inference |
| #612 | 3855677486 | Implemented here: accepted PLaMo2 activation
spellings are explicit |
| #613 | 3855845678 | Implemented here: canonical issue URL |
| #613 | 3855845757 | Implemented here: Mamba-1 function-registration
docs |
| #613 | 3855845806 | Implemented here: test expects the canonical issue
URL |
| #614 | 3855988545 | Implemented here: removed stale Nemotron-H
divergence comments |
| #614 | 3855988597 | Already fixed by #628: zero-head geometry raises
actionable `ValueError` |
| #615 | 3856082290 | Already fixed by #626: dense GraniteHybrid bias
closure |
| #618 | 3856729330 | Implemented here: required routes filter ORT GenAI
evidence |
| #618 | 3856729409 | Implemented here: env-selected runtime version is
authoritative |
| #618 | 3856729490 | Stale/N/A: PR-description-only matrix claim;
repository workflow claims one pinned version |
| #618 | 3856729563 | Implemented here: schema tail restored to normal
indentation |
| #619 | 3856342484 | Already fixed on base: Kimi Linear uses
`/issues/605` |
| #619 | 3856342532 | Already fixed by #628: config rejects convolution
kernels below 2 |
| #619 | 3856342580 | Already fixed by #628: GGUF contract rejects
convolution kernels below 2 |
| #620 | 3855717041 | Already fixed by #627: tied LM-head-only
checkpoints are retained |
| #621 | 3856722324 | Already fixed by #628: Kimi-K3 required metadata
is complete |
| #623 | 3855931210 | N/A to current main: comment belongs to open,
unmerged #623 |
| #623 | 3855939302 | N/A to current main: comment belongs to open,
unmerged #623 |
| #623 | 3855939358 | N/A to current main: comment belongs to open,
unmerged #623 |
| #623 | 3855939394 | N/A to current main: comment belongs to open,
unmerged #623 |
| #624 | 3856777152 | Implemented here: runtime compatibility reuses the
emitted model type |
| #625 | 3857049733 | Implemented here: unsupported header reports both
endian candidates |
| #629 | 3857313182 | Newer post-audit behavioral sidecar-symlink
cleanup finding; unchanged |
| #629 | 3857313251 | Newer post-audit cross-platform path-safety
finding; unchanged |

## Validation

- affected GGUF/package/ORT GenAI/model/schema tests: 1,040 passed
- broad non-integration suite: 7,851 passed, 56 skipped, 1 subtest
passed
- generated GGUF docs checks: 7 passed
- initialized `lintrunner`; full lint/format passed
- GPT-5.6 Sol medium review: one integrity finding fixed; re-review
found no significant issues

Signed-off-by: Justin Chu <justinchuby@users.noreply.github.com>
Co-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants