tqcq/llama.cpp
0
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-07-29 13:43:38 -04:00
Code Issues Packages Projects Releases Wiki Activity

Init - first pass.

Browse Source
This commit is contained in:
Vaibhavs10
2025-06-17 15:03:34 +02:00
parent e434e69183
commit 024bd29445
6 changed files with 133 additions and 9 deletions
Download Patch File Download Diff File Expand all files Collapse all files

17
gguf-py/gguf/constants.py
View File

@@ -346,6 +346,7 @@ class MODEL_ARCH(IntEnum):
BAILINGMOE = auto()
DOTS1 = auto()
ARCEE = auto()
SMOLLM3 = auto()
class VISION_PROJECTOR_TYPE(IntEnum):
@@ -629,6 +630,7 @@ MODEL_ARCH_NAMES: dict[MODEL_ARCH, str] = {
MODEL_ARCH.BAILINGMOE: "bailingmoe",
MODEL_ARCH.DOTS1: "dots1",
MODEL_ARCH.ARCEE: "arcee",
MODEL_ARCH.SMOLLM3: "smollm3",
}
VISION_PROJECTOR_TYPE_NAMES: dict[VISION_PROJECTOR_TYPE, str] = {
@@ -2101,6 +2103,21 @@ MODEL_TENSORS: dict[MODEL_ARCH, list[MODEL_TENSOR]] = {
MODEL_TENSOR.FFN_DOWN,
MODEL_TENSOR.FFN_UP,
],
MODEL_ARCH.SMOLLM3: [
MODEL_TENSOR.TOKEN_EMBD,
MODEL_TENSOR.OUTPUT_NORM,
MODEL_TENSOR.OUTPUT,
MODEL_TENSOR.ROPE_FREQS,
MODEL_TENSOR.ATTN_NORM,
MODEL_TENSOR.ATTN_Q,
MODEL_TENSOR.ATTN_K,
MODEL_TENSOR.ATTN_V,
MODEL_TENSOR.ATTN_OUT,
MODEL_TENSOR.ATTN_ROT_EMBD,
MODEL_TENSOR.FFN_GATE,
MODEL_TENSOR.FFN_DOWN,
MODEL_TENSOR.FFN_UP,
],
# TODO
}