File tree Expand file tree Collapse file tree 1 file changed +3
-3
lines changed Expand file tree Collapse file tree 1 file changed +3
-3
lines changed Original file line number Diff line number Diff line change @@ -320,7 +320,7 @@ bool llm_graph_input_attn_kv_unified::can_reuse(const llm_graph_params & params)
320
320
bool res = true ;
321
321
322
322
res &= self_k_idxs->ne [0 ] == params.ubatch .n_tokens ;
323
- res &= self_v_idxs->ne [0 ] == params.ubatch .n_tokens ;
323
+ // res &= self_v_idxs->ne[0] == params.ubatch.n_tokens; // TODO: need to move this to the unified cache and check there
324
324
325
325
res &= self_kq_mask->ne [0 ] == mctx->get_n_kv ();
326
326
res &= self_kq_mask->ne [1 ] == GGML_PAD (params.ubatch .n_tokens , GGML_KQ_MASK_PAD);
@@ -350,10 +350,10 @@ bool llm_graph_input_attn_kv_unified_iswa::can_reuse(const llm_graph_params & pa
350
350
bool res = true ;
351
351
352
352
res &= self_k_idxs->ne [0 ] == params.ubatch .n_tokens ;
353
- res &= self_v_idxs->ne [0 ] == params.ubatch .n_tokens ;
353
+ // res &= self_v_idxs->ne[0] == params.ubatch.n_tokens; // TODO: need to move this to the unified cache and check there
354
354
355
355
res &= self_k_idxs_swa->ne [0 ] == params.ubatch .n_tokens ;
356
- res &= self_v_idxs_swa->ne [0 ] == params.ubatch .n_tokens ;
356
+ // res &= self_v_idxs_swa->ne[0] == params.ubatch.n_tokens; // TODO: need to move this to the unified cache and check there
357
357
358
358
res &= self_kq_mask->ne [0 ] == mctx->get_base ()->get_n_kv ();
359
359
res &= self_kq_mask->ne [1 ] == GGML_PAD (params.ubatch .n_tokens , GGML_KQ_MASK_PAD);
You can’t perform that action at this time.
0 commit comments