Skip to content

Commit aba6ed2

Browse files
ngu3ngu3
authored
Merge branch 'main' into fix/wan-14b-perf-recipes
2 parents f0a51e8 + e2f2496 commit aba6ed2

53 files changed

Lines changed: 5570 additions & 162 deletions

File tree

Some content is hidden

Large Commits have some content hidden by default. Use the searchbox below for content that may be hidden.

‎.main.commit‎

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -1 +1 @@
1-
07046cc6b8b7479ba05e6b47ed25f3f8e6634573
1+
3e95097b4a9ab6ea6b2a39eb2b2b24e4a6dfa54a

‎3rdparty/Megatron-LM‎

Submodule Megatron-LM updated 97 files

‎README.md‎

Lines changed: 3 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -11,6 +11,8 @@
1111
</div>
1212

1313
## 📣 News
14+
- [10/15/2026] [**Muse-Spark-Open-816B-A42B**](examples/models/muse_spark_open/README.md) is now supported! Megatron Bridge adds checkpoint conversion, inference, and GB200 training recipes for the 816B-total / 42B-active vision-and-audio-language MoE model: pretraining, full SFT, long-context SFT at 32K and 64K with context parallelism, LoRA, and image-text SFT and LoRA, with HybridEP MoE dispatch. See the [examples README](examples/models/muse_spark_open/README.md) and the [verification card](examples/model_verification_cards/muse-spark-open-816b-a42b/card.yaml) for details.
15+
1416
- [08/19/2026] **Megatron Bridge 0.6.0 released!** Highlights include GLM-5.2 with cuDNN fused DSA for 128K THD-packed context-parallel training, new LLM/VLM/diffusion models, more reliable VLM data and sequence packing, improved checkpoint and LoRA workflows, unified training and conversion launchers, and refreshed performance recipes. Huge thanks to our community contributors! See the [full release notes](https://github.com/NVIDIA-NeMo/Megatron-Bridge/releases/tag/v0.6.0).
1517

1618
- [08/11/2026] **NVIDIA Nemotron 3.5 Lightning is released!** Read the technical blog [here](https://developer.nvidia.com/blog/nvidia-nemotron-3-5-lightning-delivers-fast-accurate-specialized-task-execution-for-long-running-agents/). See the [Nemotron 3.5 Lightning README](https://github.com/NVIDIA-NeMo/Megatron-Bridge/blob/main/examples/models/nemotron/nemotron_3_5_lightning/README.md) for details.
@@ -242,6 +244,7 @@ Megatron Bridge provides out-of-the-box bridges and training recipes for a wide
242244
| [**Xiaomi-MiMo**](docs/models/mimo/index.md) | Xiaomi-MiMo, MiMo-V2-Flash |
243245
| [**Moonlight**](docs/models/moonlight/index.md) | Moonlight |
244246
| [**Muse Glimmer**](docs/models/muse_glimmer/muse-glimmer.md) | Muse Glimmer 30B |
247+
| [**Muse Spark Open**](docs/models/muse_spark_open/muse-spark-open.md) | Muse-Spark-Open-816B-A42B |
245248
| [**Nemotron**](docs/models/nemotron/index.md) | Nemotron-3 Nano, Nemotron-3 Super, Nemotron-3 Ultra, Nemotron-3 Nano Omni |
246249
| [**OLMoE**](docs/models/olmoe/index.md) | OLMoE |
247250
| [**Qwen**](docs/models/qwen/index.md) | Qwen2 / Qwen2.5, Qwen3, Qwen3-MoE, Qwen3 Next, Qwen3.5 (dense/MoE), Qwen2.5-VL, Qwen3-VL, Qwen3.5-VL, Qwen3.6-VL, Qwen2 Audio, Qwen2.5-Omni, Qwen3-Omni, Qwen3-ASR |

‎docs/_static/model-verification/catalog-v1.json‎

Lines changed: 356 additions & 4 deletions
Large diffs are not rendered by default.

‎docs/conf.py‎

Lines changed: 0 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -57,9 +57,6 @@
5757
"_build",
5858
"Thumbs.db",
5959
".DS_Store",
60-
# Keep these guides in GitHub, but omit them from the published HTML.
61-
"models/nemotron/nemotron3.5-super-vl.md",
62-
"models/nemotron/nemotron3.5-super-vl-text-only.md",
6360
"skills/linting-and-formatting/SKILL.md",
6461
"skills/parity-testing/SKILL.md",
6562
"skills/pr-review/SKILL.md",

‎docs/fern/docs.yml‎

Lines changed: 0 additions & 20 deletions
Original file line numberDiff line numberDiff line change
@@ -21,23 +21,3 @@ versions:
2121
- display-name: nightly
2222
path: versions/nightly.yml
2323
slug: nightly
24-
25-
redirects:
26-
# Unpublished guides remain in GitHub; send bookmarks to Supported Models.
27-
# The catalog landing page exists in nightly, not in the frozen default version.
28-
- source: /nemo/megatron-bridge/supported-models/nemotron/nemotron-3-5-super-vl
29-
destination: /nemo/megatron-bridge/nightly/supported-models/supported-models
30-
- source: /nemo/megatron-bridge/nightly/supported-models/nemotron/nemotron-3-5-super-vl
31-
destination: /nemo/megatron-bridge/nightly/supported-models/supported-models
32-
- source: /nemo/megatron-bridge/supported-models/nemotron/nemotron-3-5-super-text-only
33-
destination: /nemo/megatron-bridge/nightly/supported-models/supported-models
34-
- source: /nemo/megatron-bridge/nightly/supported-models/nemotron/nemotron-3-5-super-text-only
35-
destination: /nemo/megatron-bridge/nightly/supported-models/supported-models
36-
- source: /nemo/megatron-bridge/models/nemotron/nemotron3.5-super-vl.html
37-
destination: /nemo/megatron-bridge/nightly/supported-models/supported-models
38-
- source: /nemo/megatron-bridge/nightly/models/nemotron/nemotron3.5-super-vl.html
39-
destination: /nemo/megatron-bridge/nightly/supported-models/supported-models
40-
- source: /nemo/megatron-bridge/models/nemotron/nemotron3.5-super-vl-text-only.html
41-
destination: /nemo/megatron-bridge/nightly/supported-models/supported-models
42-
- source: /nemo/megatron-bridge/nightly/models/nemotron/nemotron3.5-super-vl-text-only.html
43-
destination: /nemo/megatron-bridge/nightly/supported-models/supported-models

‎docs/fern/versions/nightly.yml‎

Lines changed: 4 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -124,7 +124,10 @@ navigation:
124124
path: ./nightly/pages/models/nemotron/nemotron3-ultra.mdx
125125
- page: Nemotron 3.5 Lightning
126126
path: ./nightly/pages/models/nemotron/nemotron3.5-lightning.mdx
127-
# Super VL guides remain in GitHub, but are excluded from publication.
127+
- page: Nemotron 3.5 Super VL
128+
path: ./nightly/pages/models/nemotron/nemotron3.5-super-vl.mdx
129+
- page: Nemotron-3.5 Super Text-only
130+
path: ./nightly/pages/models/nemotron/nemotron3.5-super-vl-text-only.mdx
128131
section: Nemotron
129132
- contents:
130133
- page: OLMoE

‎docs/fern/versions/nightly/pages/models/README.mdx‎

Lines changed: 15 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -4,7 +4,7 @@
44

55
Choose a model to open its interactive import/export and training configurations. Every option comes directly from an authoritative YAML verification card; independent fields are never combined into synthetic commands.
66

7-
Current inventory: **26 model cards** and **365 concrete configurations**.
7+
Current inventory: **29 model cards** and **413 concrete configurations**.
88

99
Each model page lets you select import/export, pretraining, benchmarking, SFT, LoRA, or long-context SFT and immediately see the exact command and expected result for the recorded precision and GPU.
1010

@@ -72,13 +72,17 @@ Choose a model to inspect its recorded import/export, training, and precision co
7272
<section class="verification-provider-card">
7373
<header class="verification-provider-heading">
7474
<strong>Meta</strong>
75-
<span>1 model</span>
75+
<span>2 models</span>
7676
</header>
7777
<div class="verification-provider-models">
7878
<a class="verification-model-link" href="muse_glimmer/muse-glimmer#verified-muse-glimmer-30b">
7979
<strong title="meta-models/Muse-Glimmer-30B">Muse-Glimmer-30B</strong>
8080
<span class="verification-status verification-status--partial" title="Partial">◐ Partial</span>
8181
</a>
82+
<a class="verification-model-link" href="muse_spark_open/muse-spark-open#verified-muse-spark-open-816b-a42b">
83+
<strong title="meta-models/Muse-Spark-Open-816B-A42B">Muse-Spark-Open-816B-A42B</strong>
84+
<span class="verification-status verification-status--partial" title="Partial">◐ Partial</span>
85+
</a>
8286
</div>
8387
</section>
8488
<section class="verification-provider-card">
@@ -112,7 +116,7 @@ Choose a model to inspect its recorded import/export, training, and precision co
112116
<section class="verification-provider-card">
113117
<header class="verification-provider-heading">
114118
<strong>NVIDIA</strong>
115-
<span>6 models</span>
119+
<span>8 models</span>
116120
</header>
117121
<div class="verification-provider-models">
118122
<a class="verification-model-link" href="nemotron/nemotron3-nano#verified-nemotron-3-nano-30b-a3b">
@@ -141,6 +145,14 @@ Choose a model to inspect its recorded import/export, training, and precision co
141145
<strong title="nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16">NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16</strong>
142146
<span class="verification-status verification-status--verified" title="Verified">✓ Verified</span>
143147
</a>
148+
<a class="verification-model-link" href="nemotron/nemotron3.5-super-vl#verified-nemotron-3.5-super-vl-120b-a12b">
149+
<strong title="nvidia/NVIDIA-Nemotron-3.5-Super-VL-120B-A12B-BF16">NVIDIA-Nemotron-3.5-Super-VL-120B-A12B-BF16</strong>
150+
<span class="verification-status verification-status--partial" title="Partial">◐ Partial</span>
151+
</a>
152+
<a class="verification-model-link" href="nemotron/nemotron3.5-super-vl-text-only#verified-nemotron-3.5-super-vl-120b-a12b-text-only">
153+
<strong title="nvidia/NVIDIA-Nemotron-3.5-Super-VL-120B-A12B-BF16">NVIDIA-Nemotron-3.5-Super-VL-120B-A12B-BF16</strong>
154+
<span class="verification-status verification-status--partial" title="Partial">◐ Partial</span>
155+
</a>
144156
</div>
145157
</section>
146158
<section class="verification-provider-card">

0 commit comments

Comments
 (0)