Skip to content

self.gate dtype update for GLM-4.5 #22203

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 5 commits into from
Aug 5, 2025
Merged

self.gate dtype update for GLM-4.5 #22203

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
5 commits
Select commit Hold shift + click to select a range
e2b1ab9
fuse fp32
zRzRzRzRzRzRzR Aug 3, 2025
54b2e8b
Merge branch 'vllm-project:main' into glm-45
zRzRzRzRzRzRzR Aug 4, 2025
e124e40
Update glm4_moe.py
zRzRzRzRzRzRzR Aug 4, 2025
04c219f
Merge branch 'vllm-project:main' into glm-45
zRzRzRzRzRzRzR Aug 4, 2025
b1c9528
rename
zRzRzRzRzRzRzR Aug 4, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion docs/models/supported_models.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -606,7 +606,7 @@ These models primarily accept the [`LLM.generate`](./generative_models.md#llmgen
| `GLM4VForCausalLM`<sup>^</sup> | GLM-4V | T + I | `zai-org/glm-4v-9b`, `zai-org/cogagent-9b-20241220`, etc. | ✅︎ | ✅︎ | ✅︎ |
| `Glm4vForConditionalGeneration` | GLM-4.1V-Thinking | T + I<sup>E+</sup> + V<sup>E+</sup> | `zai-org/GLM-4.1V-9B-Thinking`, etc. | ✅︎ | ✅︎ | ✅︎ |
| `Glm4MoeForCausalLM` | GLM-4.5 | T + I<sup>E+</sup> + V<sup>E+</sup> | `zai-org/GLM-4.5`, etc. | ✅︎ | ✅︎ | ✅︎ |
| `Glm4v_moeForConditionalGeneration` | GLM-4.5V | T + I<sup>E+</sup> + V<sup>E+</sup> | `zai-org/GLM-4.5V-Air`, etc. | ✅︎ | ✅︎ | ✅︎ |
| `Glm4v_moeForConditionalGeneration` | GLM-4.5V | T + I<sup>E+</sup> + V<sup>E+</sup> | `zai-org/GLM-4.5V`, etc. | ✅︎ | ✅︎ | ✅︎ |
| `GraniteSpeechForConditionalGeneration` | Granite Speech | T + A | `ibm-granite/granite-speech-3.3-8b` | ✅︎ | ✅︎ | ✅︎ |
| `H2OVLChatModel` | H2OVL | T + I<sup>E+</sup> | `h2oai/h2ovl-mississippi-800m`, `h2oai/h2ovl-mississippi-2b`, etc. | | ✅︎ | ✅︎ |
| `Idefics3ForConditionalGeneration` | Idefics3 | T + I | `HuggingFaceM4/Idefics3-8B-Llama3`, etc. | ✅︎ | | ✅︎ |
Expand Down
2 changes: 1 addition & 1 deletion tests/models/registry.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -383,7 +383,7 @@ def check_available_online(
trust_remote_code=True,
hf_overrides={"architectures": ["GLM4VForCausalLM"]}), # noqa: E501
"Glm4vForConditionalGeneration": _HfExamplesInfo("zai-org/GLM-4.1V-9B-Thinking"), # noqa: E501
"Glm4v_moeForConditionalGeneration": _HfExamplesInfo("zai-org/GLM-4.5V-Air",
"Glm4v_moeForConditionalGeneration": _HfExamplesInfo("zai-org/GLM-4.5V",
is_available_online=False), # noqa: E501
"H2OVLChatModel": _HfExamplesInfo("h2oai/h2ovl-mississippi-800m",
trust_remote_code=True,
Expand Down
3 changes: 2 additions & 1 deletion vllm/model_executor/models/glm4_moe.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -123,6 +123,7 @@ def __init__(
config.n_routed_experts,
bias=False,
quant_config=None,
params_dtype=torch.float32,
prefix=f"{prefix}.gate")

self.gate.e_score_correction_bias = nn.Parameter(
Expand Down Expand Up @@ -180,7 +181,7 @@ def forward(self, hidden_states: torch.Tensor) -> torch.Tensor:

if self.n_shared_experts is not None:
shared_output = self.shared_experts(hidden_states)
router_logits, _ = self.gate(hidden_states)
router_logits, _ = self.gate(hidden_states.to(dtype=torch.float32))
Copy link
Contributor

Choose a reason for hiding this comment

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

high

The explicit cast hidden_states.to(dtype=torch.float32) is redundant and introduces an unnecessary memory copy, which can negatively impact performance.

Since self.gate.weight is already of dtype=torch.float32 (due to the change in __init__), torch.nn.functional.linear (which is called internally by ColumnParallelLinear) will automatically perform the matrix multiplication in float32 by upcasting the hidden_states tensor. This implicit type promotion is more efficient than an explicit cast.

Removing the explicit cast will rely on this standard PyTorch behavior and avoid the overhead, while still achieving the goal of performing the gate computation in float32.

Suggested change
router_logits, _ = self.gate(hidden_states.to(dtype=torch.float32))
router_logits, _ = self.gate(hidden_states)

final_hidden_states = self.experts(
hidden_states=hidden_states,
router_logits=router_logits) * self.routed_scaling_factor
Expand Down