Add MixtralForCausalLM in Turbomind#4623
Open
43758726 wants to merge 1 commit into
Open
Conversation
43758726
changed the title
Comment on lines
+43
to
+50
| self._n_experts = getattr(cfg, 'num_experts', 0) | ||
|
|
||
| if self._n_experts > 0: | ||
| self._moe_cfg = make_moe_config( | ||
| cfg, | ||
| experts_per_token=cfg.num_experts_per_tok) | ||
| self._moe_cfg.expert_num = self._n_experts | ||
|
|
Comment on lines
+107
to
+110
| experts = ModuleListBuilder(ModuleListConfig(), self._ctx) | ||
| for e in range(self.cfg.num_experts): | ||
| experts[e] = self.ffn(pfx + 'experts' + e, is_expert=True) | ||
| m.experts = experts.build() |
Comment on lines
+121
to
+124
| if self._n_experts: | ||
| d.moe_ffn = self.moe(p + 'block_sparse_moe') | ||
| else: | ||
| d.feed_forward = self.ffn(p + 'block_sparse_moe') |
lvhan028
reviewed
May 26, 2026
Comment on lines
+51
to
+53
| # ------------------------------------------------------------------ | ||
| # model() — walks full hierarchy | ||
| # ------------------------------------------------------------------ |
lvhan028
reviewed
May 26, 2026
| def __init__(self, cfg: MixtralConfig, *, resolver): | ||
| super().__init__(cfg, resolver=resolver) | ||
|
|
||
| self._attn_cfg = make_attention_config(cfg) |
Collaborator
There was a problem hiding this comment.
to suit make_attention_config rather than changing make_attention_config
lvhan028
reviewed
May 26, 2026
Comment on lines
+64
to
+66
| lm_pfx = (pfx + 'model.embed_tokens' | ||
| if self.cfg.tie_word_embeddings | ||
| else pfx + 'lm_head') |
Collaborator
There was a problem hiding this comment.
mixtral has no tie_word_embeddings
lvhan028
reviewed
May 26, 2026
| self._ffn_cfg = make_ffn_config(cfg, | ||
| act_type=_act_type_id('silu')) | ||
|
|
||
| self._n_experts = getattr(cfg, 'num_experts', 0) |
Collaborator
There was a problem hiding this comment.
Overusing getattr(cfg, 'num_experts', 0) and similar patterns makes the code overly defensive and obscures the intent. If num_experts is a required field, just use cfg.num_experts and let the missing attribute fail explicitly. If a default value is truly appropriate, please double-check its correctness. Otherwise, reject this cargo-cult style of getattr usage.
lvhan028
reviewed
May 26, 2026
|
|
||
| self._n_experts = getattr(cfg, 'num_experts', 0) | ||
|
|
||
| if self._n_experts > 0: |
Collaborator
There was a problem hiding this comment.
We can assert self._n_experts > 0. if self._n_experts >0 can be removed
lvhan028
reviewed
May 26, 2026
Comment on lines
+121
to
+124
| if self._n_experts: | ||
| d.moe_ffn = self.moe(p + 'block_sparse_moe') | ||
| else: | ||
| d.feed_forward = self.ffn(p + 'block_sparse_moe') |
Collaborator
There was a problem hiding this comment.
Is there "mlp" layer in mixtral?
lvhan028
reviewed
May 26, 2026
| hidden_dim = cfg.hidden_size | ||
| head_num = cfg.num_attention_heads | ||
| head_dim = head_dim if head_dim is not None else getattr(cfg, 'head_dim', hidden_dim // head_num) | ||
| cfg_head_dim = getattr(cfg, 'head_dim', None) |
Collaborator
There was a problem hiding this comment.
May not change this function
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Thanks for your contribution and we appreciate it a lot. The following instructions would make your pull request more healthy and more easily receiving feedbacks. If you do not understand some items, don't worry, just make the pull request and seek help from maintainers.
Motivation
Please describe the motivation of this PR and the goal you want to achieve through this PR.
Modification
Please briefly describe what modification is made in this PR.
BC-breaking (Optional)
Does the modification introduce changes that break the backward-compatibility of the downstream repositories?
If so, please describe how it breaks the compatibility and how the downstream projects should modify their code to keep compatibility with this PR.
Use cases (Optional)
If this PR introduces a new feature, it is better to list some use cases here, and update the documentation.
Checklist