minke.yu/sglang
1
0
Fork 0
Code Issues Pull Requests Actions 12 Packages Projects Releases Wiki Activity

[Doc] Rename A5 product name (#31998)

Browse Source
This commit is contained in:
axx-ty911
2026-07-22 15:09:05 +08:00
committed by GitHub
parent ddaf430e6c
commit 7dcf3f7cbf
1 changed files with 11 additions and 11 deletions
Download Patch File Download Diff File Expand all files Collapse all files
docs_new/docs/hardware-platforms/ascend-npus/ascend_npu_quantization.mdx
+11 -11
View File
@@ -15,7 +15,7 @@ SGLang supports **mix-bits** quantization (independently defines and loads each
<th>Layer type</th>
<th>A2 Supported</th>
<th>A3 Supported</th>
<th>A5 Supported</th>
<th>Ascend 950 Products Supported</th>
<th>Diffusion models</th>
</tr>
</thead>
@@ -119,7 +119,7 @@ SGLang supports **mix-bits** quantization (independently defines and loads each
<th>Layer type</th>
<th>A2 Supported</th>
<th>A3 Supported</th>
<th>A5 Supported</th>
<th>Ascend 950 Products Supported</th>
</tr>
</thead>
<tbody>
@@ -155,7 +155,7 @@ GPTQ on Ascend support
<th>Layer type</th>
<th>A2 Supported</th>
<th>A3 Supported</th>
<th>A5 Supported</th>
<th>Ascend 950 Products Supported</th>
</tr>
</thead>
<tbody>
@@ -198,7 +198,7 @@ GPTQ on Ascend support
<th>Layer type</th>
<th>A2 Supported</th>
<th>A3 Supported</th>
<th>A5 Supported</th>
<th>Ascend 950 Products Supported</th>
</tr>
</thead>
<tbody>
@@ -241,7 +241,7 @@ Compressed-tensors (LLM Compressor) on Ascend support:
<th>Layer type</th>
<th>A2 Supported</th>
<th>A3 Supported</th>
<th>A5 Supported</th>
<th>Ascend 950 Products Supported</th>
</tr>
</thead>
<tbody>
@@ -284,7 +284,7 @@ Compressed-tensors (LLM Compressor) on Ascend support:
<th>Layer type</th>
<th>A2 Supported</th>
<th>A3 Supported</th>
<th>A5 Supported</th>
<th>Ascend 950 Products Supported</th>
</tr>
</thead>
<tbody>
@@ -335,7 +335,7 @@ python3 -m sglang.launch_server \
**MXFP8 for LLM dense models (e.g., Qwen3 / Qwen3.5):**
LLM dense W8A8 MXFP8 Linear support on Ascend was added in [PR #22352](https://github.com/sgl-project/sglang/pull/22352). Requires Ascend A5 series or newer (`npu_dynamic_mx_quant` is not available on A2 / A3).
LLM dense W8A8 MXFP8 Linear support on Ascend was added in [PR #22352](https://github.com/sgl-project/sglang/pull/22352). Requires Ascend 950 Products or newer (`npu_dynamic_mx_quant` is not available on A2 / A3).
- Online MXFP8 quantization (BF16/FP16 weights → MXFP8 at load time):
@@ -365,7 +365,7 @@ python3 -m sglang.launch_server \
**MXFP4 W4A8 for LLM dense models (e.g., Qwen3 / Qwen3.5):**
LLM dense W4A8 (MXFP4 4-bit weights + MXFP8 8-bit activations) Linear support was added in [PR #23650](https://github.com/sgl-project/sglang/pull/23650). Requires Ascend A5 series or newer.
LLM dense W4A8 (MXFP4 4-bit weights + MXFP8 8-bit activations) Linear support was added in [PR #23650](https://github.com/sgl-project/sglang/pull/23650). Requires Ascend 950 Products or newer.
- Online W4A8 quantization (BF16/FP16 weights → MXFP4 at load time):
@@ -410,9 +410,9 @@ python3 -m sglang.launch_server \
## Diffusion Model Quantization on Ascend NPU
SGLang-Diffusion supports MXFP8 online and offline quantization for diffusion models (such as Wan2.2) on Ascend NPUs. MXFP8 requires A5; the ModelSlim W8A8/W4A4 schemes work on A2/A3.
SGLang-Diffusion supports MXFP8 online and offline quantization for diffusion models (such as Wan2.2) on Ascend NPUs. MXFP8 requires Ascend 950 Products; the ModelSlim W8A8/W4A4 schemes work on A2/A3.
**Requirements for MXFP8:** CANN ≥ 8.0.RC3, Ascend A5
**Requirements for MXFP8:** CANN ≥ 8.0.RC3, Ascend 950 Products
<table>
<thead>
@@ -422,7 +422,7 @@ SGLang-Diffusion supports MXFP8 online and offline quantization for diffusion mo
<th>Scheme class</th>
<th>Mode</th>
<th>A2/A3 Supported</th>
<th>A5 Supported</th>
<th>Ascend 950 Products Supported</th>
<th>Trigger</th>
</tr>
</thead>