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

more cleaning on python code

Browse Source
This commit is contained in:
younesbelkada
2025-07-03 18:09:30 +04:00
parent fdd5cff4ba
commit 14c37ec047
5 changed files with 204 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

40
gguf-py/gguf/constants.py
View File

@@ -172,6 +172,7 @@ class Keys:
TIME_STEP_RANK = "{arch}.ssm.time_step_rank"
GROUP_COUNT = "{arch}.ssm.group_count"
DT_B_C_RMS = "{arch}.ssm.dt_b_c_rms"
HEAD_DIM = "{arch}.ssm.head_dim"
class WKV:
HEAD_SIZE = "{arch}.wkv.head_size"
@@ -288,6 +289,7 @@ class MODEL_ARCH(IntEnum):
LLAMA4 = auto()
DECI = auto()
FALCON = auto()
FALCON_H1 = auto()
BAICHUAN = auto()
GROK = auto()
GPT2 = auto()
@@ -525,6 +527,7 @@ class MODEL_TENSOR(IntEnum):
POSNET_ATTN_K = auto()
POSNET_ATTN_V = auto()
POSNET_ATTN_OUT = auto()
SSM_MUP_VEC = auto()
# vision
V_MMPROJ = auto()
V_MMPROJ_FC = auto()
@@ -660,6 +663,7 @@ MODEL_ARCH_NAMES: dict[MODEL_ARCH, str] = {
MODEL_ARCH.DOTS1: "dots1",
MODEL_ARCH.ARCEE: "arcee",
MODEL_ARCH.ERNIE4_5: "ernie4_5",
MODEL_ARCH.FALCON_H1: "falcon_h1",
}
VISION_PROJECTOR_TYPE_NAMES: dict[VISION_PROJECTOR_TYPE, str] = {
@@ -736,6 +740,7 @@ TENSOR_NAMES: dict[MODEL_TENSOR, str] = {
MODEL_TENSOR.SSM_D: "blk.{bid}.ssm_d",
MODEL_TENSOR.SSM_NORM: "blk.{bid}.ssm_norm",
MODEL_TENSOR.SSM_OUT: "blk.{bid}.ssm_out",
MODEL_TENSOR.SSM_MUP_VEC: "blk.{bid}.ssm_mup_vec",
MODEL_TENSOR.TIME_MIX_W0: "blk.{bid}.time_mix_w0",
MODEL_TENSOR.TIME_MIX_W1: "blk.{bid}.time_mix_w1",
MODEL_TENSOR.TIME_MIX_W2: "blk.{bid}.time_mix_w2",
@@ -2211,6 +2216,41 @@ MODEL_TENSORS: dict[MODEL_ARCH, list[MODEL_TENSOR]] = {
MODEL_TENSOR.FFN_DOWN,
MODEL_TENSOR.FFN_UP,
],
MODEL_ARCH.FALCON_H1: [
# Token embedding
MODEL_TENSOR.TOKEN_EMBD,
# Input layernorm
MODEL_TENSOR.ATTN_NORM,
# Attention components
MODEL_TENSOR.ATTN_Q, # Query projection
MODEL_TENSOR.ATTN_K, # Key projection
MODEL_TENSOR.ATTN_V, # Value projection
MODEL_TENSOR.ATTN_OUT, # Output projection
# SSM components (Mamba2 specific)
MODEL_TENSOR.SSM_MUP_VEC, # Mup vector
MODEL_TENSOR.SSM_IN, # Input projection for SSM
MODEL_TENSOR.SSM_CONV1D, # Convolution layer
MODEL_TENSOR.SSM_DT, # Delta time projection
MODEL_TENSOR.SSM_A, # A parameter (log form)
MODEL_TENSOR.SSM_D, # D parameter
MODEL_TENSOR.SSM_NORM, # Normalization in SSM
MODEL_TENSOR.SSM_OUT, # Output projection
# Pre-feedforward layernorm
MODEL_TENSOR.FFN_PRE_NORM,
# Feed-forward network components
MODEL_TENSOR.FFN_GATE, # Gate projection (SwiGLU)
MODEL_TENSOR.FFN_DOWN, # Down projection
MODEL_TENSOR.FFN_UP, # Up projection
# Post-feedforward layernorm
MODEL_TENSOR.OUTPUT_NORM, # Final layer norm
MODEL_TENSOR.OUTPUT, # Output projection (lm_head)
],
# TODO
}

9
gguf-py/gguf/gguf_writer.py
View File

@@ -867,6 +867,15 @@ class GGUFWriter:
def add_ssm_dt_b_c_rms(self, value: bool) -> None:
self.add_bool(Keys.SSM.DT_B_C_RMS.format(arch=self.arch), value)
def add_ssm_head_dim(self, value: int) -> None:
self.add_uint32(Keys.SSM.HEAD_DIM.format(arch=self.arch), value)
def add_attn_head_count(self, count: int) -> None:
self.add_uint32(Keys.Attention.HEAD_COUNT.format(arch=self.arch), count)
def add_key_value_head_count(self, count: int) -> None:
self.add_uint32(Keys.Attention.HEAD_COUNT_KV.format(arch=self.arch), count)
def add_tokenizer_model(self, model: str) -> None:
self.add_string(Keys.Tokenizer.MODEL, model)

16
gguf-py/gguf/tensor_mapping.py
View File

@@ -286,12 +286,14 @@ class TensorNameMap:
# Post feed-forward norm
MODEL_TENSOR.FFN_PRE_NORM: (
"model.layers.{bid}.pre_feedforward_layernorm", # gemma2
"model.layers.{bid}.pre_ff_layernorm.weight",
),
# Post feed-forward norm
MODEL_TENSOR.FFN_POST_NORM: (
"model.layers.{bid}.post_feedforward_layernorm", # gemma2 olmo2
"model.layers.{bid}.post_mlp_layernorm", # glm-4-0414
"model.layers.{bid}.feed_forward.up_proj",
),
MODEL_TENSOR.FFN_GATE_INP: (
@@ -362,6 +364,7 @@ class TensorNameMap:
"model.layers.{bid}.mlp.shared_expert.up_proj", # qwen2moe
"model.layers.{bid}.mlp.shared_experts.up_proj", # deepseek deepseek2
"model.layers.{bid}.feed_forward.shared_expert.up_proj", # llama4
"model.layers.{bid}.feed_forward.down_proj",
),
# AWQ-activation gate
@@ -547,11 +550,13 @@ class TensorNameMap:
MODEL_TENSOR.SSM_IN: (
"model.layers.{bid}.in_proj",
"backbone.layers.{bid}.mixer.in_proj",
"model.layers.{bid}.mamba.in_proj",
),
MODEL_TENSOR.SSM_CONV1D: (
"model.layers.{bid}.conv1d",
"backbone.layers.{bid}.mixer.conv1d",
"model.layers.{bid}.mamba.conv1d",
),
MODEL_TENSOR.SSM_X: (
@@ -562,16 +567,19 @@ class TensorNameMap:
MODEL_TENSOR.SSM_DT: (
"model.layers.{bid}.dt_proj",
"backbone.layers.{bid}.mixer.dt_proj",
"model.layers.{bid}.mamba.dt_proj",
),
MODEL_TENSOR.SSM_A: (
"model.layers.{bid}.A_log",
"backbone.layers.{bid}.mixer.A_log",
"model.layers.{bid}.mamba.A_log",
),
MODEL_TENSOR.SSM_D: (
"model.layers.{bid}.D",
"backbone.layers.{bid}.mixer.D",
"model.layers.{bid}.mamba.D",
),
MODEL_TENSOR.SSM_NORM: (
@@ -1168,6 +1176,14 @@ class TensorNameMap:
"resampler.attn.out_proj",
),
MODEL_TENSOR.SSM_MUP_VEC: (
"model.layers.{bid}.mamba.mup_vector", # falcon-h1
),
MODEL_TENSOR.SSM_NORM: (
"model.layers.{bid}.mamba.norm",
),
MODEL_TENSOR.V_RESMPL_KV: (
"resampler.kv_proj",
),