Skip to content

fix: resolve LLM norm tensor names by architecture #1570

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
leejet merged 1 commit into from
May 27, 2026
+24 −19
Merged

fix: resolve LLM norm tensor names by architecture #1570

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
39 changes: 22 additions & 17 deletions src/llm.hpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -948,15 +948,20 @@ namespace LLM {
protected:
LLMArch arch;
int sliding_attention;
bool has_post_attention_norm;
bool has_post_ffw_norm;
std::string post_attention_norm_name;
std::string pre_ffw_norm_name;
std::string post_ffw_norm_name;

public:
TransformerBlock(const LLMParams& params, int layer_index)
: arch(params.arch),
sliding_attention(0),
has_post_attention_norm(params.arch == LLMArch::GEMMA3_12B),
has_post_ffw_norm(params.arch == LLMArch::GEMMA3_12B) {
sliding_attention(0) {
if (params.arch == LLMArch::GEMMA3_12B) {
post_attention_norm_name = "post_attention_norm";
post_ffw_norm_name = "post_ffw_norm";
}
pre_ffw_norm_name = params.arch == LLMArch::GPT_OSS_20B ? "post_attention_norm" : "post_attention_layernorm";

blocks["self_attn"] = std::make_shared<Attention>(params);
if (params.arch == LLMArch::GPT_OSS_20B) {
blocks["mlp"] = std::make_shared<GPTOSSMLP>(params);
Expand All @@ -966,13 +971,13 @@ namespace LLM {
false,
params.mlp_activation);
}
blocks["input_layernorm"] = std::make_shared<LLMRMSNorm>(params.hidden_size, params.rms_norm_eps, params.rms_norm_add);
blocks["post_attention_layernorm"] = std::make_shared<LLMRMSNorm>(params.hidden_size, params.rms_norm_eps, params.rms_norm_add);
if (has_post_attention_norm) {
blocks["post_attention_norm"] = std::make_shared<LLMRMSNorm>(params.hidden_size, params.rms_norm_eps, params.rms_norm_add);
blocks["input_layernorm"] = std::make_shared<LLMRMSNorm>(params.hidden_size, params.rms_norm_eps, params.rms_norm_add);
blocks[pre_ffw_norm_name] = std::make_shared<LLMRMSNorm>(params.hidden_size, params.rms_norm_eps, params.rms_norm_add);
if (!post_attention_norm_name.empty()) {
blocks[post_attention_norm_name] = std::make_shared<LLMRMSNorm>(params.hidden_size, params.rms_norm_eps, params.rms_norm_add);
}
if (has_post_ffw_norm) {
blocks["post_ffw_norm"] = std::make_shared<LLMRMSNorm>(params.hidden_size, params.rms_norm_eps, params.rms_norm_add);
if (!post_ffw_norm_name.empty()) {
blocks[post_ffw_norm_name] = std::make_shared<LLMRMSNorm>(params.hidden_size, params.rms_norm_eps, params.rms_norm_add);
}
if (!params.sliding_attention.empty()) {
sliding_attention = params.sliding_attention[layer_index % params.sliding_attention.size()];
Expand All @@ -987,14 +992,14 @@ namespace LLM {
// x: [N, n_token, hidden_size]
auto self_attn = std::dynamic_pointer_cast<Attention>(blocks["self_attn"]);
auto input_layernorm = std::dynamic_pointer_cast<LLMRMSNorm>(blocks["input_layernorm"]);
auto post_attention_layernorm = std::dynamic_pointer_cast<LLMRMSNorm>(blocks["post_attention_layernorm"]);
auto pre_ffw_norm = std::dynamic_pointer_cast<LLMRMSNorm>(blocks[pre_ffw_norm_name]);
std::shared_ptr<LLMRMSNorm> post_attention_norm = nullptr;
std::shared_ptr<LLMRMSNorm> post_ffw_norm = nullptr;
if (has_post_attention_norm) {
post_attention_norm = std::dynamic_pointer_cast<LLMRMSNorm>(blocks["post_attention_norm"]);
if (!post_attention_norm_name.empty()) {
post_attention_norm = std::dynamic_pointer_cast<LLMRMSNorm>(blocks[post_attention_norm_name]);
}
if (has_post_ffw_norm) {
post_ffw_norm = std::dynamic_pointer_cast<LLMRMSNorm>(blocks["post_ffw_norm"]);
if (!post_ffw_norm_name.empty()) {
post_ffw_norm = std::dynamic_pointer_cast<LLMRMSNorm>(blocks[post_ffw_norm_name]);
}
ggml_tensor* block_attention_mask = attention_mask;
int rope_index = 0;
Expand All @@ -1012,7 +1017,7 @@ namespace LLM {
x = ggml_add_inplace(ctx->ggml_ctx, x, residual);

residual = x;
x = post_attention_layernorm->forward(ctx, x);
x = pre_ffw_norm->forward(ctx, x);
if (arch == LLMArch::GPT_OSS_20B) {
auto mlp = std::dynamic_pointer_cast<GPTOSSMLP>(blocks["mlp"]);
x = mlp->forward(ctx, x);
Expand Down
4 changes: 2 additions & 2 deletions src/name_conversion.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -138,15 +138,15 @@ std::string convert_cond_stage_model_name(std::string name, std::string prefix)
{"attn_k_norm.", "self_attn.k_norm."},
{"attn_output.", "self_attn.o_proj."},
{"attn_norm.", "input_layernorm."},
{"attn_post_norm.", "post_attention_layernorm."},
{"post_attention_norm.", "post_attention_layernorm."},
{"attn_post_norm.", "post_attention_norm."},
{"ffn_gate_inp.", "mlp.router."},
{"ffn_gate_exps.", "mlp.experts.gate_proj."},
{"ffn_up_exps.", "mlp.experts.up_proj."},
{"ffn_down_exps.", "mlp.experts.down_proj."},
{"ffn_down.", "mlp.down_proj."},
{"ffn_gate.", "mlp.gate_proj."},
{"ffn_up.", "mlp.up_proj."},
{"ffn_post_norm.", "post_ffw_norm."},
{"ffn_norm.", "post_attention_layernorm."},
{"output_norm.", "model.norm."},
};
Expand Down
Loading