Skip to content

Rename Indexers to follow existing conventions #2

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
timkhronos merged 1 commit into from
Jun 26, 2026
Merged

Rename Indexers to follow existing conventions #2

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
22 changes: 10 additions & 12 deletions gguf-py/gguf/constants.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -742,10 +742,9 @@ class MODEL_TENSOR(IntEnum):
INDEXER_PROJ = auto()
INDEXER_ATTN_K = auto()
INDEXER_ATTN_Q_B = auto()
INDEX_Q_PROJ = auto()
INDEX_K_PROJ = auto()
INDEX_Q_NORM = auto()
INDEX_K_NORM = auto()
INDEXER_Q_PROJ = auto()
INDEXER_K_PROJ = auto()
INDEXER_Q_NORM = auto()
# vision
V_MMPROJ = auto()
V_MMPROJ_FC = auto()
Expand Down Expand Up @@ -1304,10 +1303,9 @@ class MODEL_TENSOR(IntEnum):
MODEL_TENSOR.INDEXER_PROJ: "blk.{bid}.indexer.proj",
MODEL_TENSOR.INDEXER_ATTN_K: "blk.{bid}.indexer.attn_k",
MODEL_TENSOR.INDEXER_ATTN_Q_B: "blk.{bid}.indexer.attn_q_b",
MODEL_TENSOR.INDEX_Q_PROJ: "blk.{bid}.index_q_proj",
MODEL_TENSOR.INDEX_K_PROJ: "blk.{bid}.index_k_proj",
MODEL_TENSOR.INDEX_Q_NORM: "blk.{bid}.index_q_norm",
MODEL_TENSOR.INDEX_K_NORM: "blk.{bid}.index_k_norm",
MODEL_TENSOR.INDEXER_Q_PROJ: "blk.{bid}.indexer.q_proj",
MODEL_TENSOR.INDEXER_K_PROJ: "blk.{bid}.indexer.k_proj",
MODEL_TENSOR.INDEXER_Q_NORM: "blk.{bid}.indexer.q_norm",
# vision
MODEL_TENSOR.V_MMPROJ: "mm.{bid}",
MODEL_TENSOR.V_MMPROJ_FC: "mm.model.fc",
Expand Down Expand Up @@ -4040,10 +4038,10 @@ class MODEL_TENSOR(IntEnum):
MODEL_TENSOR.FFN_GATE,
MODEL_TENSOR.FFN_DOWN,
MODEL_TENSOR.FFN_UP,
MODEL_TENSOR.INDEX_Q_PROJ,
MODEL_TENSOR.INDEX_K_PROJ,
MODEL_TENSOR.INDEX_Q_NORM,
MODEL_TENSOR.INDEX_K_NORM,
MODEL_TENSOR.INDEXER_Q_PROJ,
MODEL_TENSOR.INDEXER_K_PROJ,
MODEL_TENSOR.INDEXER_Q_NORM,
MODEL_TENSOR.INDEXER_K_NORM,



Expand Down
16 changes: 6 additions & 10 deletions gguf-py/gguf/tensor_mapping.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1263,7 +1263,8 @@ class TensorNameMap:
),

MODEL_TENSOR.INDEXER_K_NORM: (
"model.layers.{bid}.self_attn.indexer.k_norm", # DSA
"model.layers.{bid}.self_attn.indexer.k_norm", # DSA
"model.layers.{bid}.self_attn.index_k_norm", # MSA
),

MODEL_TENSOR.INDEXER_PROJ: (
Expand All @@ -1278,22 +1279,17 @@ class TensorNameMap:
"model.layers.{bid}.self_attn.indexer.wq_b", # DSA
),

MODEL_TENSOR.INDEX_Q_PROJ: (
MODEL_TENSOR.INDEXER_Q_PROJ: (
"model.layers.{bid}.self_attn.index_q_proj", # MSA
),

MODEL_TENSOR.INDEX_K_PROJ: (
MODEL_TENSOR.INDEXER_K_PROJ: (
"model.layers.{bid}.self_attn.index_k_proj", # MSA
),

MODEL_TENSOR.INDEX_Q_NORM: (
MODEL_TENSOR.INDEXER_Q_NORM: (
"model.layers.{bid}.self_attn.index_q_norm", # MSA
),

MODEL_TENSOR.INDEX_K_NORM: (
"model.layers.{bid}.self_attn.index_k_norm", # MSA
),

),


############################################################################
Expand Down
14 changes: 6 additions & 8 deletions src/llama-arch.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -568,10 +568,9 @@ static const std::map<llm_tensor, const char *> LLM_TENSOR_NAMES = {
{ LLM_TENSOR_INDEXER_PROJ, "blk.%d.indexer.proj" },
{ LLM_TENSOR_INDEXER_ATTN_K, "blk.%d.indexer.attn_k" },
{ LLM_TENSOR_INDEXER_ATTN_Q_B, "blk.%d.indexer.attn_q_b" },
{ LLM_TENSOR_INDEX_Q_PROJ, "blk.%d.index_q_proj" },
{ LLM_TENSOR_INDEX_K_PROJ, "blk.%d.index_k_proj" },
{ LLM_TENSOR_INDEX_Q_NORM, "blk.%d.index_q_norm" },
{ LLM_TENSOR_INDEX_K_NORM, "blk.%d.index_k_norm" },
{ LLM_TENSOR_INDEXER_Q_PROJ, "blk.%d.indexer.q_proj" },
{ LLM_TENSOR_INDEXER_K_PROJ, "blk.%d.indexer.k_proj" },
{ LLM_TENSOR_INDEXER_Q_NORM, "blk.%d.indexer.q_norm" },
{ LLM_TENSOR_MASKED_EMBD_CENTROIDS, "masked_embd_centroids" },
{ LLM_TENSOR_MASKED_EMBD_ORDERING, "masked_embd_ordering" },
{ LLM_TENSOR_FC, "fc" },
Expand Down Expand Up @@ -785,10 +784,9 @@ static const std::map<llm_tensor, llm_tensor_info> LLM_TENSOR_INFOS = {
{LLM_TENSOR_INDEXER_PROJ, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}},
{LLM_TENSOR_INDEXER_ATTN_K, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}},
{LLM_TENSOR_INDEXER_ATTN_Q_B, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}},
{LLM_TENSOR_INDEX_Q_PROJ, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}},
{LLM_TENSOR_INDEX_K_PROJ, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}},
{LLM_TENSOR_INDEX_Q_NORM, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL}},
{LLM_TENSOR_INDEX_K_NORM, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL}},
{LLM_TENSOR_INDEXER_Q_PROJ, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}},
{LLM_TENSOR_INDEXER_K_PROJ, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}},
{LLM_TENSOR_INDEXER_Q_NORM, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL}},
{LLM_TENSOR_NEXTN_PROJ_PRE, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}},
{LLM_TENSOR_NEXTN_PROJ_POST, {LLM_TENSOR_LAYER_OUTPUT, GGML_OP_MUL_MAT}},
// NextN/MTP tensors are stored per-block (blk.%d.nextn.*) even though only the
Expand Down
15 changes: 7 additions & 8 deletions src/llama-arch.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -564,14 +564,13 @@ enum llm_tensor {
LLM_TENSOR_VISEXP_FFN_GATE,
LLM_TENSOR_VISEXP_FFN_DOWN,
LLM_TENSOR_VISEXP_FFN_UP,
LLM_TENSOR_INDEXER_K_NORM,
LLM_TENSOR_INDEXER_PROJ,
LLM_TENSOR_INDEXER_ATTN_K,
LLM_TENSOR_INDEXER_ATTN_Q_B,
LLM_TENSOR_INDEX_Q_PROJ,
LLM_TENSOR_INDEX_K_PROJ,
LLM_TENSOR_INDEX_Q_NORM,
LLM_TENSOR_INDEX_K_NORM,
LLM_TENSOR_INDEXER_K_NORM, //DSA, MSA
LLM_TENSOR_INDEXER_PROJ, //DSA
LLM_TENSOR_INDEXER_ATTN_K, //DSA
LLM_TENSOR_INDEXER_ATTN_Q_B, //DSA
LLM_TENSOR_INDEXER_Q_PROJ, //MSA
LLM_TENSOR_INDEXER_K_PROJ, //MSA
LLM_TENSOR_INDEXER_Q_NORM, //MSA
LLM_TENSOR_NEXTN_PROJ_PRE,
LLM_TENSOR_NEXTN_PROJ_POST,
LLM_TENSOR_NEXTN_EH_PROJ,
Expand Down
8 changes: 4 additions & 4 deletions src/models/minimax-m3.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -70,10 +70,10 @@ void llama_model_minimax_m3::load_arch_tensors(llama_model_loader &) {
layer.ffn_up_shexp = create_tensor(tn(LLM_TENSOR_FFN_UP_SHEXP, "weight", i), {n_embd, n_ff_exp * n_expert_shared}, 0);

// indexer
layer.index_q_proj = create_tensor(tn(LLM_TENSOR_INDEX_Q_PROJ, "weight", i), {n_embd, hparams.indexer_n_head * hparams.indexer_head_size}, 0);
layer.index_k_proj = create_tensor(tn(LLM_TENSOR_INDEX_K_PROJ, "weight", i), {n_embd, hparams.indexer_head_size}, 0);
layer.index_q_norm = create_tensor(tn(LLM_TENSOR_INDEX_Q_NORM, "weight", i), {hparams.indexer_head_size}, 0);
layer.index_k_norm = create_tensor(tn(LLM_TENSOR_INDEX_K_NORM, "weight", i), {hparams.indexer_head_size}, 0);
layer.index_q_proj = create_tensor(tn(LLM_TENSOR_INDEXER_Q_PROJ, "weight", i), {n_embd, hparams.indexer_n_head * hparams.indexer_head_size}, 0);
layer.index_k_proj = create_tensor(tn(LLM_TENSOR_INDEXER_K_PROJ, "weight", i), {n_embd, hparams.indexer_head_size}, 0);
layer.index_q_norm = create_tensor(tn(LLM_TENSOR_INDEXER_Q_NORM, "weight", i), {hparams.indexer_head_size}, 0);
layer.index_k_norm = create_tensor(tn(LLM_TENSOR_INDEXER_K_NORM, "weight", i), {hparams.indexer_head_size}, 0);
}
}
}
Expand Down
Loading