CS348Project/llama.cpp
5
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-10-30 08:42:00 +00:00
Code Issues Packages Projects Releases Wiki Activity

model : add grok-2 support (#15539)

Browse Source 
* add grok-2 support

* type fix

* type fix

* type fix

* "fix" vocab for invalid sequences

* fix expert tensor mapping and spaces in vocab

* add chat template

* fix norm tensor mapping

* rename layer_out_norm to ffn_post_norm

* ensure ffn_post_norm is mapped

* fix experts merging

* remove erroneous FFN_GATE entry

* concatenate split tensors and add more metadata

* process all expert layers and try cat instead of hstack

* add support for community BPE vocab

* fix expert feed forward length and ffn_down concat

* commit this too

* add ffn_up/gate/down, unsure if sequence is right

* add ffn_gate/down/up to tensor names

* correct residual moe (still not working)

* mess--

* fix embedding scale being applied twice

* add built in chat template

* change beta fast for grok if default value

* remove spm vocab in favor of community bpe vocab

* change attention temp length metadata type to integer

* update attention temp length metadata

* remove comment

* replace M_SQRT2 with std::sqrt(2)

* add yarn metadata, move defaults to hparams
This commit is contained in:
Sigbjørn Skjæret
2025-09-14 23:00:59 +02:00
committed by GitHub
parent 6c019cb04e
commit b8e09f08b9
16 changed files with 281 additions and 96 deletions
Download Patch File Download Diff File Expand all files Collapse all files

10
gguf-py/gguf/tensor_mapping.py
View File

@@ -136,6 +136,7 @@ class TensorNameMap:
"model.layers.{bid}.norm", # mamba-qbert
"backbone.layers.{bid}.norm", # mamba
"transformer.decoder_layer.{bid}.rms_norm", # Grok
"model.layers.{bid}.pre_attn_norm", # grok-2
"transformer.blocks.{bid}.norm_attn_norm.norm_1", # dbrx
"encoder.layers.{bid}.input_layernorm", # chatglm
"transformer.layers.{bid}.attn_norm", # openelm
@@ -278,6 +279,7 @@ class TensorNameMap:
"transformer.layer.{bid}.sa_layer_norm", # distillbert
"encoder.layers.{bid}.norm1", # nomic-bert
"transformer.decoder_layer.{bid}.rms_norm_1", # Grok
"model.layers.{bid}.post_attn_norm", # grok-2
"transformer.blocks.{bid}.norm_attn_norm.norm_2", # dbrx
),
@@ -313,6 +315,7 @@ class TensorNameMap:
"h.{bid}.ln_2", # gpt2
"model.layers.{bid}.ffn_norm", # internlm2
"transformer.decoder_layer.{bid}.rms_norm_2", # Grok
"model.layers.{bid}.pre_moe_norm", # grok-2
"encoder.layers.{bid}.post_attention_layernorm", # chatglm
"transformer.layers.{bid}.ffn_norm", # openelm
"model.layers.{bid}.pre_ff_layernorm", # jamba granite-hybrid
@@ -333,11 +336,12 @@ class TensorNameMap:
# Post feed-forward norm
MODEL_TENSOR.FFN_POST_NORM: (
"model.layers.{bid}.post_feedforward_layernorm", # gemma2 olmo2
"layers.{bid}.post_feedforward_layernorm", # embeddinggemma
"model.layers.{bid}.post_mlp_layernorm", # glm-4-0414
"model.layers.{bid}.post_feedforward_layernorm", # gemma2 olmo2
"layers.{bid}.post_feedforward_layernorm", # embeddinggemma
"model.layers.{bid}.post_mlp_layernorm", # glm-4-0414
"model.layers.layers.{bid}.post_mlp_norm.weight", # plamo2
"model.layers.{bid}.feed_forward.up_proj",
"model.layers.{bid}.post_moe_norm", # grok-2
),
MODEL_TENSOR.FFN_GATE_INP: (