MCPcopy Index your code
hub / github.com/abetlen/llama-cpp-python / main

Function main

examples/server/server.py:16352–16460  ·  view source on GitHub ↗
()

Source from the content-addressed store, hash-verified

16350
16351
16352def main() -> None:
16353 parser = argparse.ArgumentParser()
16354 parser.add_argument("-C", "--config-file", required=True)
16355 args = parser.parse_args()
16356 config = ConfigFile.load(args.config_file)
16357 model_path = config.model.resolve_model_path()
16358 loras = [
16359 Model.LoraAdapter(path=lora.resolve_path(), scale=lora.scale)
16360 for lora in config.model.loras
16361 ]
16362 model = Model(
16363 model_path=model_path,
16364 model_alias=config.model.alias,
16365 chat_template=config.model.chat_template,
16366 loras=loras,
16367 n_gpu_layers=config.model.n_gpu_layers,
16368 split_mode=config.model.split_mode,
16369 main_gpu=config.model.main_gpu,
16370 tensor_split=config.model.tensor_split,
16371 vocab_only=config.model.vocab_only,
16372 use_mmap=config.model.use_mmap,
16373 use_mlock=config.model.use_mlock,
16374 kv_overrides=config.model.kv_overrides,
16375 n_ctx=config.model.n_ctx,
16376 n_batch=config.model.n_batch,
16377 n_ubatch=config.model.n_ubatch,
16378 n_seq_max=config.model.n_seq_max,
16379 n_threads=config.model.threads,
16380 n_threads_batch=config.model.threads_batch,
16381 rope_scaling_type=config.model.rope_scaling_type,
16382 pooling_type=config.model.pooling_type,
16383 attention_type=config.model.attention_type,
16384 embedding=config.model.embedding,
16385 rope_freq_base=config.model.rope_freq_base,
16386 rope_freq_scale=config.model.rope_freq_scale,
16387 yarn_ext_factor=config.model.yarn_ext_factor,
16388 yarn_attn_factor=config.model.yarn_attn_factor,
16389 yarn_beta_fast=config.model.yarn_beta_fast,
16390 yarn_beta_slow=config.model.yarn_beta_slow,
16391 yarn_orig_ctx=config.model.yarn_orig_ctx,
16392 offload_kqv=config.model.offload_kqv,
16393 flash_attn=config.model.flash_attn,
16394 op_offload=config.model.op_offload,
16395 swa_full=config.model.swa_full,
16396 no_perf=config.model.no_perf,
16397 type_k=config.model.type_k,
16398 type_v=config.model.type_v,
16399 kv_unified=config.model.kv_unified,
16400 max_seq_len=config.model.max_seq_len,
16401 max_output_tokens=config.model.max_output_tokens,
16402 draft_model=config.model.draft_model,
16403 draft_model_path=config.model.resolve_draft_model_path(),
16404 draft_model_num_pred_tokens=config.model.draft_model_num_pred_tokens,
16405 draft_model_max_ngram_size=config.model.draft_model_max_ngram_size,
16406 draft_model_top_k=config.model.draft_model_top_k,
16407 draft_model_p_min=config.model.draft_model_p_min,
16408 draft_model_max_batch_size=config.model.draft_model_max_batch_size,
16409 draft_model_threads=config.model.draft_model_threads,

Callers 1

server.pyFile · 0.70

Calls 14

ModelClass · 0.85
MTMDEmbeddingCacheClass · 0.85
MTMDProcessorClass · 0.85
SequenceDiskCacheClass · 0.85
CompletionSchedulerClass · 0.85
CompletionServiceClass · 0.85
resolve_pathMethod · 0.80
resolve_mmproj_pathMethod · 0.80
loadMethod · 0.45
resolve_model_pathMethod · 0.45

Tested by

no test coverage detected

Used in the wild real call sites across dependent graphs

searching dependent graphs…