ikawrakow/ik_llama.cpp
1
0
Fork 0
mirror of https://github.com/ikawrakow/ik_llama.cpp.git synced 2026-04-21 15:09:40 +00:00
Code Issues Packages Projects Releases Wiki Activity

Add command line argument for fused delta net

Browse Source
This commit is contained in:
Kawrakow
2026-02-24 05:40:26 +00:00
parent a350f1b96f
commit 28b31a66b2
6 changed files with 18 additions and 6 deletions
Download Patch File Download Diff File Expand all files Collapse all files

7
src/llama-delta-net.cpp
View File

@@ -679,10 +679,9 @@ ggml_tensor * delta_net::build_layer_attn_linear_core(ggml_context * ctx0, ggml_
GGML_ASSERT(identity != nullptr);
GGML_ASSERT(diag_mask != nullptr);
attn_out = n_tok == 1
//? build_delta_net_autoregressive(ctx0, q_conv, k_conv, v_conv, gate, beta, state, il, cb)
? build_fused_delta_net(ctx0, q_conv, k_conv, v_conv, gate, beta, state, il, cb)
: build_delta_net_chunking(ctx0, q_conv, k_conv, v_conv, gate, beta, state, causal_mask, identity, diag_mask, il, cb);
attn_out = n_tok == 1 ? lctx.cparams.fused_delta_net ? build_fused_delta_net(ctx0, q_conv, k_conv, v_conv, gate, beta, state, il, cb)
: build_delta_net_autoregressive(ctx0, q_conv, k_conv, v_conv, gate, beta, state, il, cb)
: build_delta_net_chunking(ctx0, q_conv, k_conv, v_conv, gate, beta, state, causal_mask, identity, diag_mask, il, cb);
ggml_tensor * output = attn_out.first;
ggml_tensor * new_state = attn_out.second;
cb(output, "attn_output", il);