CS348Project/llama.cpp
5
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-11-04 09:32:00 +00:00
Code Issues Packages Projects Releases Wiki Activity

remove todo

Browse Source
This commit is contained in:
younesbelkada
2025-07-08 10:44:59 +04:00
parent 2dee7cf964
commit a846d02327
2 changed files with 2 additions and 6 deletions
Download Patch File Download Diff File Expand all files Collapse all files

5
gguf-py/gguf/tensor_mapping.py
View File

@@ -583,6 +583,7 @@ class TensorNameMap:
), ),
MODEL_TENSOR.SSM_NORM: ( MODEL_TENSOR.SSM_NORM: (
"model.layers.{bid}.mamba.norm", # falcon-h1
"backbone.layers.{bid}.mixer.norm", # mamba2 "backbone.layers.{bid}.mixer.norm", # mamba2
), ),
@@ -1177,10 +1178,6 @@ class TensorNameMap:
"resampler.attn.out_proj", "resampler.attn.out_proj",
), ),
MODEL_TENSOR.SSM_NORM: (
"model.layers.{bid}.mamba.norm",
),
MODEL_TENSOR.V_RESMPL_KV: ( MODEL_TENSOR.V_RESMPL_KV: (
"resampler.kv_proj", "resampler.kv_proj",
), ),

3
src/llama-arch.cpp
View File

@@ -1955,8 +1955,7 @@ bool llm_arch_is_recurrent(const llm_arch & arch) {
} }
bool llm_arch_is_hybrid(const llm_arch & arch) { bool llm_arch_is_hybrid(const llm_arch & arch) {
// TODO: There are currently no hybrid models! Once there are, this will be // List all mamba-attention hybrid models here
// the place to identify them
switch (arch) { switch (arch) {
case LLM_ARCH_FALCON_H1: case LLM_ARCH_FALCON_H1:
return true; return true;