Read the equation
In DeepseekV3MoE.moe_infer, this statement defines or updates gatherd_idxs[s:s + k]. This statement evaluates the named projection, normalization or numeric transform with the source arguments. Its learned parameters and exact input shape are defined by the surrounding module.
Terms & definitions
The recovered passage does not supply a separate symbol table. Read the source context for the notation used in this version; familiar letters may carry different meanings in other builds.
How it works
Read gatherd_idxs[s:s + k] as the quantity being defined or updated by this relation.
Uses a remainder or modular operation, retaining the numeric representation used in the source.
Use this version’s variable definitions and update order. The same symbol may be assigned a different meaning in another build.
Assumptions & limits
Source recovery establishes that the expression appears in this record. Its mathematical validity, implementation behavior, and interpretation must be assessed separately. See the linked source and research notes for the checks available for this version.
Original context
Read the surrounding source passage
dtype=np.int32)
s = 0
for i, k in enumerate(tokens_per_expert_group.cpu().numpy()):
gatherd_idxs[s:s + k] = i % self.experts_per_rank
s += k
gatherd_idxs = gatherd_idxs.argsort()
sorted_tokens = gathered_tokens[gatherd_idxs]Exact original notation
gatherd_idxs[s:s + k] = i % self.experts_per_rank
Related builds
Sources & provenance
gatherd_idxs[s : s + k] = i % self.experts_per_rank