fix: add graph cuts for MiniMax-H3 text conditioning (#1900)

This commit is contained in:
assouan authored and GitHub committed 2026-09-24 00:03:29 +08:00
1 parent 2dc7f5408a
commit 36746936c0
2 files changed
+15 -8

No files matched your search

+7 -7
View File
@@ -178,13 +178,13 @@ namespace Qwen {
auto h = std::dynamic_pointer_cast<Linear>(blocks[name])->forward(ctx, x);
return ggml_reshape_4d(ctx->ggml_ctx, h, dim_head, heads, x->ne[1], x->ne[2]);
};
auto q = project("to_q");
auto k = project("to_k");
auto v = project("to_v");
q = std::dynamic_pointer_cast<RMSNorm>(blocks["norm_q"])->forward(ctx, q);
k = std::dynamic_pointer_cast<RMSNorm>(blocks["norm_k"])->forward(ctx, k);
q = Rope::apply_rope(ctx->ggml_ctx, q, pe);
k = Rope::apply_rope(ctx->ggml_ctx, k, pe);
auto q = project("to_q");
auto k = project("to_k");
auto v = project("to_v");
q = std::dynamic_pointer_cast<RMSNorm>(blocks["norm_q"])->forward(ctx, q);
k = std::dynamic_pointer_cast<RMSNorm>(blocks["norm_k"])->forward(ctx, k);
q = Rope::apply_rope(ctx->ggml_ctx, q, pe);
k = Rope::apply_rope(ctx->ggml_ctx, k, pe);
if (cache.mode == QwenImage21PrefixCache::Mode::STORE) {
auto persist = [&](ggml_tensor* tensor, int axis, const char* name) {
auto part = ggml_ext_slice(ctx->ggml_ctx, tensor, axis, 0, cache.prefix_length);