CS348Project/llama.cpp
5
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-10-31 08:51:55 +00:00
Code Issues Packages Projects Releases Wiki Activity

convert.py : n_head_kv optional and .gguf file extension

Browse Source
This commit is contained in:
klosax
2023-08-17 17:20:36 +02:00
committed by GitHub
parent dd016cc246
commit d646c4efce
1 changed files with 23 additions and 18 deletions
Download Patch File Download Diff File Expand all files Collapse all files

41
convert.py
View File

@@ -150,15 +150,20 @@ class Params:
def loadHFTransformerJson(model: 'LazyModel', config_path: 'Path') -> 'Params':
config = json.load(open(config_path))
n_vocab = config["vocab_size"];
n_embd = config["hidden_size"];
n_layer = config["num_hidden_layers"];
n_ff = config["intermediate_size"];
n_head = config["num_attention_heads"];
n_head_kv = config["num_key_value_heads"];
f_norm_eps = config["rms_norm_eps"];
n_vocab = config["vocab_size"]
n_embd = config["hidden_size"]
n_layer = config["num_hidden_layers"]
n_ff = config["intermediate_size"]
n_head = config["num_attention_heads"]
n_mult = Params.find_n_mult(n_ff, n_embd);
if "num_key_value_heads" in config:
n_head_kv = config["num_key_value_heads"]
else:
n_head_kv = None
f_norm_eps = config["rms_norm_eps"]
n_mult = Params.find_n_mult(n_ff, n_embd)
if "max_sequence_length" in config:
n_ctx = config["max_sequence_length"]
@@ -186,15 +191,15 @@ class Params:
def loadOriginalParamsJson(model: 'LazyModel', config_path: 'Path') -> 'Params':
config = json.load(open(config_path))
n_vocab = config["vocab_size"];
n_embd = config["dim"];
n_layer = config["n_layers"];
n_mult = config["multiple_of"];
n_vocab = config["vocab_size"]
n_embd = config["dim"]
n_layer = config["n_layers"]
n_mult = config["multiple_of"]
n_ctx = 2048 if config["norm_eps"] == 1e-06 else 4096 # hack to determine LLaMA v1 vs v2
n_ff = -1;
n_head = config["n_heads"];
n_head_kv = config["n_kv_heads"] if "n_kv_heads" in config else n_head;
f_norm_eps = config["norm_eps"];
n_ff = -1
n_head = config["n_heads"]
n_head_kv = config["n_kv_heads"] if "n_kv_heads" in config else n_head
f_norm_eps = config["norm_eps"]
if n_vocab == -1:
n_vocab = model["tok_embeddings.weight"].shape[0]
@@ -714,7 +719,7 @@ class OutputFile:
self.gguf.add_feed_forward_length (params.n_ff)
self.gguf.add_rope_dimension_count(params.n_embd // params.n_head)
self.gguf.add_head_count (params.n_head)
self.gguf.add_head_count_kv (params.n_head_kv)
if params.n_head_kv is not None: self.gguf.add_head_count_kv(params.n_head_kv)
self.gguf.add_layer_norm_rms_eps (params.f_norm_eps)
def add_meta_vocab(self, vocab: Vocab) -> None:
@@ -934,7 +939,7 @@ def default_outfile(model_paths: List[Path], file_type: GGMLFileType) -> Path:
GGMLFileType.AllF32: "f32",
GGMLFileType.MostlyF16: "f16",
}[file_type]
ret = model_paths[0].parent / f"ggml-model-{namestr}.bin"
ret = model_paths[0].parent / f"ggml-model-{namestr}.gguf"
if ret in model_paths:
sys.stderr.write(
f"Error: Default output path ({ret}) would overwrite the input. "