[New model][Multimodal] Add DeepSeek-V4-Flash-Vision-Exp support (#54566)

Signed-off-by: Isotr0py <Isotr0py@outlook.com>
This commit is contained in:
Isotr0py
2026-09-02 11:09:09 -07:00
committed by GitHub
parent 488e6fd53c
commit 1356635d83
37 changed files with 2918 additions and 152 deletions
+2 -1
View File
@@ -20,7 +20,8 @@ STABLE_TORCH_LIBRARY_FRAGMENT(_moe_C, m) {
"topk_softplus_sqrt(Tensor! topk_weights, Tensor! topk_indices, Tensor! "
"token_expert_indices, Tensor gating_output, bool renormalize, float "
"routed_scaling_factor, Tensor? "
"bias, Tensor? input_ids, Tensor? tid2eid, Tensor? is_padding) -> ()");
"bias, Tensor? input_ids, Tensor? tid2eid, Tensor? is_padding, "
"Tensor? bias_vl=None, int image_sentinel_lo=0) -> ()");
// Calculate the result of moe by summing up the partial results
// from all selected experts. topk_ids/expert_map are optional and, when