File tree Expand file tree Collapse file tree 4 files changed +8
-8
lines changed Expand file tree Collapse file tree 4 files changed +8
-8
lines changed Original file line number Diff line number Diff line change @@ -233,7 +233,7 @@ def __init__(
233
233
rope_scaling_type : Optional [int ] = llama_cpp .LLAMA_ROPE_SCALING_UNSPECIFIED ,
234
234
rope_freq_base : float = 0.0 ,
235
235
rope_freq_scale : float = 0.0 ,
236
- yarn_ext_factor : float = float ( "nan" ) ,
236
+ yarn_ext_factor : float = - 1.0 ,
237
237
yarn_attn_factor : float = 1.0 ,
238
238
yarn_beta_fast : float = 32.0 ,
239
239
yarn_beta_slow : float = 1.0 ,
Original file line number Diff line number Diff line change @@ -315,11 +315,11 @@ class llama_model_params(Structure):
315
315
316
316
317
317
# struct llama_context_params {
318
- # uint32_t seed; // RNG seed, -1 for random
319
- # uint32_t n_ctx; // text context, 0 = from model
320
- # uint32_t n_batch; // prompt processing maximum batch size
321
- # uint32_t n_threads; // number of threads to use for generation
322
- # uint32_t n_threads_batch; // number of threads to use for batch processing
318
+ # uint32_t seed; // RNG seed, -1 for random
319
+ # uint32_t n_ctx; // text context, 0 = from model
320
+ # uint32_t n_batch; // prompt processing maximum batch size
321
+ # uint32_t n_threads; // number of threads to use for generation
322
+ # uint32_t n_threads_batch; // number of threads to use for batch processing
323
323
# int8_t rope_scaling_type; // RoPE scaling type, from `enum llama_rope_scaling_type`
324
324
325
325
# // ref: https://github.com/ggerganov/llama.cpp/pull/2054
Original file line number Diff line number Diff line change @@ -93,7 +93,7 @@ class Settings(BaseSettings):
93
93
default = 0.0 , description = "RoPE frequency scaling factor"
94
94
)
95
95
yarn_ext_factor : float = Field (
96
- default = float ( "nan" )
96
+ default = - 1.0
97
97
)
98
98
yarn_attn_factor : float = Field (
99
99
default = 1.0
You can’t perform that action at this time.
0 commit comments