diff options
| author | Georgi Gerganov <ggerganov@gmail.com> | 2024-01-12 13:01:56 +0200 |
|---|---|---|
| committer | GitHub <noreply@github.com> | 2024-01-12 13:01:56 +0200 |
| commit | f445c0e68cf8e1faca0b2aa8dfb9d48231cec301 (patch) | |
| tree | a230e263424ba71153764e21fef7b92f46bd9fd9 /gguf-py/gguf/tensor_mapping.py | |
| parent | 326b418b59b6d48d854c4461a2303e8ac0a311e6 (diff) | |
llama : fix llm_build_k_shift to use correct n_rot (#4889)
* llama : fix llm_build_k_shift to use correct n_rot
ggml-ci
* llama : always use hparams.n_rot for ggml_rope_custom
ggml-ci
* convert : fix persimmon conversion to write correct n_rot
Diffstat (limited to 'gguf-py/gguf/tensor_mapping.py')
| -rw-r--r-- | gguf-py/gguf/tensor_mapping.py | 7 |
1 files changed, 7 insertions, 0 deletions
diff --git a/gguf-py/gguf/tensor_mapping.py b/gguf-py/gguf/tensor_mapping.py index 80c1d544..24a08903 100644 --- a/gguf-py/gguf/tensor_mapping.py +++ b/gguf-py/gguf/tensor_mapping.py @@ -57,6 +57,7 @@ class TensorNameMap: "transformer.norm_f", # mpt "ln_f", # refact bloom qwen gpt2 "language_model.encoder.final_layernorm", # persimmon + "model.final_layernorm", # persimmon "lm_head.ln", # phi2 ), @@ -98,6 +99,7 @@ class TensorNameMap: "transformer.h.{bid}.self_attention.query_key_value", # falcon "h.{bid}.self_attention.query_key_value", # bloom "language_model.encoder.layers.{bid}.self_attention.query_key_value", # persimmon + "model.layers.{bid}.self_attn.query_key_value", # persimmon "h.{bid}.attn.c_attn", # gpt2 "transformer.h.{bid}.mixer.Wqkv", # phi2 ), @@ -141,6 +143,7 @@ class TensorNameMap: "encoder.layer.{bid}.attention.output.dense", # bert "transformer.h.{bid}.attn.out_proj", # gpt-j "language_model.encoder.layers.{bid}.self_attention.dense", # persimmon + "model.layers.{bid}.self_attn.dense", # persimmon "h.{bid}.attn.c_proj", # gpt2 "transformer.h.{bid}.mixer.out_proj", # phi2 "model.layers.layers.{bid}.self_attn.o_proj", # plamo @@ -184,6 +187,7 @@ class TensorNameMap: "encoder.layer.{bid}.intermediate.dense", # bert "transformer.h.{bid}.mlp.fc_in", # gpt-j "language_model.encoder.layers.{bid}.mlp.dense_h_to_4h", # persimmon + "model.layers.{bid}.mlp.dense_h_to_4h", # persimmon "transformer.h.{bid}.mlp.w1", # qwen "h.{bid}.mlp.c_fc", # gpt2 "transformer.h.{bid}.mlp.fc1", # phi2 @@ -225,6 +229,7 @@ class TensorNameMap: "encoder.layer.{bid}.output.dense", # bert "transformer.h.{bid}.mlp.fc_out", # gpt-j "language_model.encoder.layers.{bid}.mlp.dense_4h_to_h", # persimmon + "model.layers.{bid}.mlp.dense_4h_to_h", # persimmon "h.{bid}.mlp.c_proj", # gpt2 "transformer.h.{bid}.mlp.fc2", # phi2 "model.layers.layers.{bid}.mlp.down_proj", # plamo @@ -237,10 +242,12 @@ class TensorNameMap: MODEL_TENSOR.ATTN_Q_NORM: ( "language_model.encoder.layers.{bid}.self_attention.q_layernorm", + "model.layers.{bid}.self_attn.q_layernorm", # persimmon ), MODEL_TENSOR.ATTN_K_NORM: ( "language_model.encoder.layers.{bid}.self_attention.k_layernorm", + "model.layers.{bid}.self_attn.k_layernorm", # persimmon ), MODEL_TENSOR.ROPE_FREQS: ( |