Skip to content

convert : proper tensor name mapping for llama4 #12870

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
merged 3 commits into from
Apr 11, 2025
Merged

convert : proper tensor name mapping for llama4 #12870

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
b14a449
Llama-4 mapping
danielhanchen Apr 8, 2025
8ed60e3
Merge pull request #20 from danielhanchen/patch-1
ngxson Apr 10, 2025
5ff43f8
remove hacky renaming
ngxson Apr 10, 2025
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
4 changes: 0 additions & 4 deletions convert_hf_to_gguf.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1788,10 +1788,6 @@ def set_gguf_parameters(self):
self.gguf_writer.add_expert_feed_forward_length(self.hparams["intermediate_size_moe"])

def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None):
name = name.replace("language_model.", "")
name = name.replace("feed_forward.", "mlp.") # a bit hacky for now
name = name.replace(".router.weight", ".gate.weight") # a bit hacky for now

# split the gate_up into gate and up
if "gate_up_proj" in name:
name_up = name.replace("gate_up_proj", "up_proj.weight")
Expand Down
19 changes: 19 additions & 0 deletions gguf-py/gguf/tensor_mapping.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -30,6 +30,7 @@ class TensorNameMap:
"rwkv.embeddings", # rwkv6
"model.embeddings", # rwkv7
"model.word_embeddings", # bailingmoe
"language_model.model.embed_tokens", # llama4
),

# Token type embeddings
Expand Down Expand Up @@ -67,6 +68,7 @@ class TensorNameMap:
"output_layer", # chatglm
"head", # rwkv
"head.out", # wavtokenizer
"language_model.lm_head", # llama4
),

# Output norm
Expand All @@ -89,6 +91,7 @@ class TensorNameMap:
"rwkv.ln_out", # rwkv6
"model.ln_out", # rwkv7
"backbone.final_layer_norm", # wavtokenizer
"language_model.model.norm", # llama4
),

# Rope frequencies
Expand Down Expand Up @@ -130,6 +133,7 @@ class TensorNameMap:
"transformer.layers.{bid}.attn_norm", # openelm
"rwkv.blocks.{bid}.ln1", # rwkv6
"model.layers.{bid}.ln1", # rwkv7
"language_model.model.layers.{bid}.input_layernorm", # llama4
),

# Attention norm 2
Expand Down Expand Up @@ -169,6 +173,7 @@ class TensorNameMap:
"model.layers.{bid}.attention.wq", # internlm2
"transformer.decoder_layer.{bid}.multi_head_attention.query",# Grok
"transformer.h.{bid}.attn.attention.q_proj", # exaone
"language_model.model.layers.{bid}.self_attn.q_proj", # llama4
),

# Attention key
Expand All @@ -183,6 +188,7 @@ class TensorNameMap:
"model.layers.{bid}.attention.wk", # internlm2
"transformer.decoder_layer.{bid}.multi_head_attention.key",# Grok
"transformer.h.{bid}.attn.attention.k_proj", # exaone
"language_model.model.layers.{bid}.self_attn.k_proj", # llama4
),

# Attention value
Expand All @@ -196,6 +202,7 @@ class TensorNameMap:
"model.layers.{bid}.attention.wv", # internlm2
"transformer.decoder_layer.{bid}.multi_head_attention.value",# Grok
"transformer.h.{bid}.attn.attention.v_proj", # exaone
"language_model.model.layers.{bid}.self_attn.v_proj", # llama4
),

# Attention output
Expand All @@ -222,6 +229,7 @@ class TensorNameMap:
"encoder.layers.{bid}.self_attention.dense", # chatglm
"transformer.layers.{bid}.attn.out_proj", # openelm
"transformer.h.{bid}.attn.attention.out_proj", # exaone
"language_model.model.layers.{bid}.self_attn.o_proj", # llama4
),

# Attention output norm
Expand Down Expand Up @@ -259,6 +267,7 @@ class TensorNameMap:
"transformer.decoder_layer.{bid}.rms_norm_2", # Grok
"encoder.layers.{bid}.post_attention_layernorm", # chatglm
"transformer.layers.{bid}.ffn_norm", # openelm
"language_model.model.layers.{bid}.post_attention_layernorm", # llama4
),

# Post feed-forward norm
Expand All @@ -278,6 +287,7 @@ class TensorNameMap:
"transformer.decoder_layer.{bid}.router", # Grok
"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
),

MODEL_TENSOR.FFN_GATE_INP_SHEXP: (
Expand Down Expand Up @@ -315,6 +325,7 @@ class TensorNameMap:
"model.layers.{bid}.residual_mlp.w3", # arctic
"encoder.layers.{bid}.mlp.dense_h_to_4h", # chatglm
"transformer.h.{bid}.mlp.c_fc_1", # exaone
"language_model.model.layers.{bid}.feed_forward.up_proj", # llama4
),

MODEL_TENSOR.FFN_UP_EXP: (
Expand All @@ -323,11 +334,13 @@ class TensorNameMap:
"transformer.blocks.{bid}.ffn.experts.mlp.v1", # dbrx
"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
),

MODEL_TENSOR.FFN_UP_SHEXP: (
"model.layers.{bid}.mlp.shared_expert.up_proj", # qwen2moe
"model.layers.{bid}.mlp.shared_experts.up_proj", # deepseek deepseek2
"language_model.model.layers.{bid}.feed_forward.shared_expert.up_proj", # llama4
),

# AWQ-activation gate
Expand All @@ -348,6 +361,7 @@ class TensorNameMap:
"transformer.h.{bid}.mlp.linear_1", # refact
"model.layers.{bid}.residual_mlp.w1", # arctic
"transformer.h.{bid}.mlp.c_fc_0", # exaone
"language_model.model.layers.{bid}.feed_forward.gate_proj", # llama4
),

MODEL_TENSOR.FFN_GATE_EXP: (
Expand All @@ -356,11 +370,13 @@ class TensorNameMap:
"transformer.blocks.{bid}.ffn.experts.mlp.w1", # dbrx
"model.layers.{bid}.mlp.experts.gate_proj", # qwen2moe olmoe (merged)
"model.layers.{bid}.block_sparse_moe.experts.w1", # phimoe (merged)
"language_model.model.layers.{bid}.feed_forward.experts.gate_proj", # llama4
),

MODEL_TENSOR.FFN_GATE_SHEXP: (
"model.layers.{bid}.mlp.shared_expert.gate_proj", # qwen2moe
"model.layers.{bid}.mlp.shared_experts.gate_proj", # deepseek deepseek2
"language_model.model.layers.{bid}.feed_forward.shared_expert.gate_proj", # llama4
),

# Feed-forward down
Expand Down Expand Up @@ -389,6 +405,7 @@ class TensorNameMap:
"encoder.layer.{bid}.mlp.down_layer", # jina-bert-v2
"encoder.layers.{bid}.mlp.dense_4h_to_h", # chatglm
"model.layers.h.{bid}.mlp.c_proj", # exaone
"language_model.model.layers.{bid}.feed_forward.down_proj", # llama4
),

MODEL_TENSOR.FFN_DOWN_EXP: (
Expand All @@ -398,11 +415,13 @@ class TensorNameMap:
"model.layers.{bid}.mlp.experts.down_proj", # qwen2moe olmoe (merged)
"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
),

MODEL_TENSOR.FFN_DOWN_SHEXP: (
"model.layers.{bid}.mlp.shared_expert.down_proj", # qwen2moe
"model.layers.{bid}.mlp.shared_experts.down_proj", # deepseek deepseek2
"language_model.model.layers.{bid}.feed_forward.shared_expert.down_proj", # llama4
),

MODEL_TENSOR.ATTN_Q_NORM: (
Expand Down