dnrops
/
llama.swift
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

remove debug printf() statements

This commit is contained in:
Alex Rozanski 2023-03-14 11:19:36 +01:00
parent 60458cc580
commit 2b27f14035
1 changed files with 0 additions and 62 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

62
llama/bridge/LlamaPredictOperation.mm
View File

@ -89,8 +89,6 @@ struct llama_model {
// load the model's weights from a file
bool llama_model_load(const std::string & fname, llama_model & model, gpt_vocab & vocab, int n_ctx) {
printf("%s: loading model from '%s' - please wait ...\n", __func__, fname.c_str());
auto fin = std::ifstream(fname, std::ios::binary);
if (!fin) {
fprintf(stderr, "%s: failed to open '%s'\n", __func__, fname.c_str());
@ -127,17 +125,6 @@ bool llama_model_load(const std::string & fname, llama_model & model, gpt_vocab
n_ff = ((2*(4*hparams.n_embd)/3 + hparams.n_mult - 1)/hparams.n_mult)*hparams.n_mult;
n_parts = LLAMA_N_PARTS.at(hparams.n_embd);
printf("%s: n_vocab = %d\n", __func__, hparams.n_vocab);
printf("%s: n_ctx = %d\n", __func__, hparams.n_ctx);
printf("%s: n_embd = %d\n", __func__, hparams.n_embd);
printf("%s: n_mult = %d\n", __func__, hparams.n_mult);
printf("%s: n_head = %d\n", __func__, hparams.n_head);
printf("%s: n_layer = %d\n", __func__, hparams.n_layer);
printf("%s: n_rot = %d\n", __func__, hparams.n_rot);
printf("%s: f16 = %d\n", __func__, hparams.f16);
printf("%s: n_ff = %d\n", __func__, n_ff);
printf("%s: n_parts = %d\n", __func__, n_parts);
}
// load vocab
@ -220,8 +207,6 @@ bool llama_model_load(const std::string & fname, llama_model & model, gpt_vocab
ctx_size += n_ctx*n_layer*n_embd*ggml_type_sizef(GGML_TYPE_F32); // memory_v
ctx_size += (5 + 10*n_layer)*256; // object overhead
printf("%s: ggml ctx size = %6.2f MB\n", __func__, ctx_size/(1024.0*1024.0));
}
// create the ggml context
@ -307,8 +292,6 @@ bool llama_model_load(const std::string & fname, llama_model & model, gpt_vocab
model.memory_v = ggml_new_tensor_1d(ctx, GGML_TYPE_F32, n_elements);
const size_t memory_size = ggml_nbytes(model.memory_k) + ggml_nbytes(model.memory_v);
printf("%s: memory_size = %8.2f MB, n_mem = %d\n", __func__, memory_size/1024.0/1024.0, n_mem);
}
const size_t file_offset = fin.tellg();
@ -326,8 +309,6 @@ bool llama_model_load(const std::string & fname, llama_model & model, gpt_vocab
fname_part += "." + std::to_string(i);
}
printf("%s: loading model part %d/%d from '%s'\n", __func__, i+1, n_parts, fname_part.c_str());
fin = std::ifstream(fname_part, std::ios::binary);
fin.seekg(file_offset);
@ -336,8 +317,6 @@ bool llama_model_load(const std::string & fname, llama_model & model, gpt_vocab
int n_tensors = 0;
size_t total_size = 0;
printf("%s: ", __func__);
while (true) {
int32_t n_dims;
int32_t length;
@ -436,7 +415,6 @@ bool llama_model_load(const std::string & fname, llama_model & model, gpt_vocab
if (0) {
static const char * ftype_str[] = { "f32", "f16", "q4_0", "q4_1", };
printf("%24s - [%5d, %5d], type = %6s, split = %d\n", name.data(), ne[0], ne[1], ftype_str[ftype], split_type);
}
size_t bpe = 0;
@ -498,17 +476,7 @@ bool llama_model_load(const std::string & fname, llama_model & model, gpt_vocab
total_size += ggml_nbytes(tensor)/n_parts;
}
//printf("%42s - [%5d, %5d], type = %6s, %6.2f MB\n", name.data(), ne[0], ne[1], ftype == 0 ? "float" : "f16", ggml_nbytes(tensor)/1024.0/1024.0);
if (++n_tensors % 8 == 0) {
printf(".");
fflush(stdout);
}
}
printf(" done\n");
printf("%s: model size = %8.2f MB / num tensors = %d\n", __func__, total_size/1024.0/1024.0, n_tensors);
}
fin.close();
@ -794,8 +762,6 @@ NSError *makeLlamaError(LlamaErrorCode errorCode, NSString *description)
ggml_time_init();
const int64_t t_main_start_us = ggml_time_us();
printf("%s: seed = %d\n", __func__, _params.seed);
std::mt19937 rng(_params.seed);
if (_params.prompt.empty()) {
_params.prompt = gpt_random_prompt(rng);
@ -841,16 +807,6 @@ NSError *makeLlamaError(LlamaErrorCode errorCode, NSString *description)
// tokenize the reverse prompt
std::vector<gpt_vocab::id> antiprompt_inp = ::llama_tokenize(vocab, _params.antiprompt, false);
printf("\n");
printf("%s: prompt: '%s'\n", __func__, _params.prompt.c_str());
printf("%s: number of tokens in prompt = %zu\n", __func__, embd_inp.size());
for (int i = 0; i < (int) embd_inp.size(); i++) {
printf("%6d -> '%s'\n", embd_inp[i], vocab.id_to_token.at(embd_inp[i]).c_str());
}
printf("\n");
printf("sampling parameters: temp = %f, top_k = %d, top_p = %f, repeat_last_n = %i, repeat_penalty = %f\n", _params.temp, _params.top_k, _params.top_p, _params.repeat_last_n, _params.repeat_penalty);
printf("\n\n");
std::vector<gpt_vocab::id> embd;
// determine the required inference memory per token:
@ -926,28 +882,10 @@ NSError *makeLlamaError(LlamaErrorCode errorCode, NSString *description)
NSString *token = [[NSString alloc] initWithCString:vocab.id_to_token[id].c_str() encoding:NSUTF8StringEncoding];
[self postEvent:[_LlamaEvent outputTokenWithToken:token]];
}
// end of text token
if (embd.back() == 2) {
printf(" [end of text]\n");
break;
}
}
[self postEvent:[_LlamaEvent completed]];
// report timing
{
const int64_t t_main_end_us = ggml_time_us();
printf("\n\n");
printf("%s: mem per token = %8zu bytes\n", __func__, mem_per_token);
printf("%s: load time = %8.2f ms\n", __func__, t_load_us/1000.0f);
printf("%s: sample time = %8.2f ms\n", __func__, t_sample_us/1000.0f);
printf("%s: predict time = %8.2f ms / %.2f ms per token\n", __func__, t_predict_us/1000.0f, t_predict_us/1000.0f/n_past);
printf("%s: total time = %8.2f ms\n", __func__, (t_main_end_us - t_main_start_us)/1000.0f);
}
ggml_free(model.ctx);
}