tqcq/llama.cpp
0
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-06-26 19:55:04 +00:00
Code Issues Packages Projects Releases Wiki Activity

model : Nomic Embed Text V2 with Mixture-of-Experts (MoE) architecture (#12466)

Browse Source 
* Nomic Embed Text V2 with Mixture-of-Experts (MoE) architecture

- Adds MoE-based embedding model supporting multilingual embeddings.
- Selects architecture variant based on hyperparameter detection (MoE layers).
- Removes unnecessary subclass initialization checks for clarity.

https://www.nomic.ai/blog/posts/nomic-embed-text-v2

Co-authored-by: Jared Van Bortel <jared@nomic.ai>

* fix tokenizer

* don't rename this tensor

---------

Co-authored-by: Jared Van Bortel <jared@nomic.ai>
This commit is contained in:
AT
2025-04-28 15:52:15 -04:00
committed by GitHub
parent eaea325324
commit 5f5e39e1ba
9 changed files with 247 additions and 110 deletions
Download Patch File Download Diff File Expand all files Collapse all files

19
gguf-py/gguf/constants.py
View File

@ -104,6 +104,7 @@ class Keys:
EXPERT_WEIGHTS_SCALE = "{arch}.expert_weights_scale"
EXPERT_WEIGHTS_NORM = "{arch}.expert_weights_norm"
EXPERT_GATING_FUNC = "{arch}.expert_gating_func"
MOE_EVERY_N_LAYERS = "{arch}.moe_every_n_layers"
POOLING_TYPE = "{arch}.pooling_type"
LOGIT_SCALE = "{arch}.logit_scale"
DECODER_START_TOKEN_ID = "{arch}.decoder_start_token_id"
@ -267,6 +268,7 @@ class MODEL_ARCH(IntEnum):
REFACT = auto()
BERT = auto()
NOMIC_BERT = auto()
NOMIC_BERT_MOE = auto()
JINA_BERT_V2 = auto()
BLOOM = auto()
STABLELM = auto()
@ -521,6 +523,7 @@ MODEL_ARCH_NAMES: dict[MODEL_ARCH, str] = {
MODEL_ARCH.REFACT: "refact",
MODEL_ARCH.BERT: "bert",
MODEL_ARCH.NOMIC_BERT: "nomic-bert",
MODEL_ARCH.NOMIC_BERT_MOE: "nomic-bert-moe",
MODEL_ARCH.JINA_BERT_V2: "jina-bert-v2",
MODEL_ARCH.BLOOM: "bloom",
MODEL_ARCH.STABLELM: "stablelm",
@ -960,6 +963,22 @@ MODEL_TENSORS: dict[MODEL_ARCH, list[MODEL_TENSOR]] = {
MODEL_TENSOR.FFN_UP,
MODEL_TENSOR.LAYER_OUT_NORM,
],
MODEL_ARCH.NOMIC_BERT_MOE: [
MODEL_TENSOR.TOKEN_EMBD,
MODEL_TENSOR.TOKEN_EMBD_NORM,
MODEL_TENSOR.TOKEN_TYPES,
MODEL_TENSOR.POS_EMBD,
MODEL_TENSOR.OUTPUT_NORM,
MODEL_TENSOR.ATTN_OUT_NORM,
MODEL_TENSOR.ATTN_QKV,
MODEL_TENSOR.ATTN_OUT,
MODEL_TENSOR.FFN_DOWN,
MODEL_TENSOR.FFN_UP,
MODEL_TENSOR.FFN_GATE_INP,
MODEL_TENSOR.FFN_DOWN_EXP,
MODEL_TENSOR.FFN_UP_EXP,
MODEL_TENSOR.LAYER_OUT_NORM,
],
MODEL_ARCH.JINA_BERT_V2: [
MODEL_TENSOR.TOKEN_EMBD,
MODEL_TENSOR.TOKEN_EMBD_NORM,

3
gguf-py/gguf/gguf_writer.py
View File

@ -728,6 +728,9 @@ class GGUFWriter:
def add_expert_gating_func(self, value: ExpertGatingFuncType) -> None:
self.add_uint32(Keys.LLM.EXPERT_GATING_FUNC.format(arch=self.arch), value.value)
def add_moe_every_n_layers(self, value: int) -> None:
self.add_uint32(Keys.LLM.MOE_EVERY_N_LAYERS.format(arch=self.arch), value)
def add_swin_norm(self, value: bool) -> None:
self.add_bool(Keys.LLM.SWIN_NORM.format(arch=self.arch), value)

4
gguf-py/gguf/tensor_mapping.py
View File

@ -290,6 +290,7 @@ class TensorNameMap:
"transformer.blocks.{bid}.ffn.router.layer", # dbrx
"model.layers.{bid}.block_sparse_moe.router.layer", # granitemoe
"language_model.model.layers.{bid}.feed_forward.router", # llama4
"encoder.layers.{bid}.mlp.router.layer", # nomic-bert-moe
),
MODEL_TENSOR.FFN_GATE_INP_SHEXP: (
@ -322,6 +323,7 @@ class TensorNameMap:
"model.layers.layers.{bid}.mlp.up_proj", # plamo
"model.layers.{bid}.feed_forward.w3", # internlm2
"encoder.layers.{bid}.mlp.fc11", # nomic-bert
"encoder.layers.{bid}.mlp.fc1", # nomic-bert-moe
"model.layers.{bid}.mlp.c_fc", # starcoder2
"encoder.layer.{bid}.mlp.gated_layers_v", # jina-bert-v2
"model.layers.{bid}.residual_mlp.w3", # arctic
@ -337,6 +339,7 @@ class TensorNameMap:
"model.layers.{bid}.mlp.experts.up_proj", # qwen2moe olmoe (merged)
"model.layers.{bid}.block_sparse_moe.experts.w3", # phimoe (merged)
"language_model.model.layers.{bid}.feed_forward.experts.up_proj", # llama4
"encoder.layers.{bid}.mlp.experts.mlp.w1", # nomic-bert-moe
),
MODEL_TENSOR.FFN_UP_SHEXP: (
@ -418,6 +421,7 @@ class TensorNameMap:
"model.layers.{bid}.block_sparse_moe.output_linear", # granitemoe
"model.layers.{bid}.block_sparse_moe.experts.w2", # phimoe (merged)
"language_model.model.layers.{bid}.feed_forward.experts.down_proj", # llama4
"encoder.layers.{bid}.mlp.experts.mlp.w2", # nomic-bert-moe
),
MODEL_TENSOR.FFN_DOWN_SHEXP: (