root/llama.cpp
1
0
Fork 0
mirror of https://github.com/ggerganov/llama.cpp.git synced 2025-01-03 23:34:35 +00:00
Code Issues Actions 7 Packages Projects Releases Wiki Activity

llama : silence errors KV cache errors

Browse Source
This commit is contained in:
Georgi Gerganov 2023-09-19 13:39:52 +03:00
parent 16090a5dde
commit d37081ae5d
No known key found for this signature in database
GPG Key ID: 449E073F9DC10735
1 changed files with 4 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

8
llama.cpp
View File

@ -1304,7 +1304,7 @@ static bool llama_kv_cache_find_slot(
} }
if (n_tested >= n_ctx) { if (n_tested >= n_ctx) {
LLAMA_LOG_ERROR("%s: failed to find a slot for %d tokens\n", __func__, n_tokens); //LLAMA_LOG_ERROR("%s: failed to find a slot for %d tokens\n", __func__, n_tokens);
return false; return false;
} }
} }
@ -7333,7 +7333,7 @@ int llama_eval(
llama_kv_cache_rm_tokens(ctx->kv_self, n_past, -1); llama_kv_cache_rm_tokens(ctx->kv_self, n_past, -1);
if (!llama_decode_internal(*ctx, llama_batch_get_one(tokens, n_tokens, n_past, 0), n_threads)) { if (!llama_decode_internal(*ctx, llama_batch_get_one(tokens, n_tokens, n_past, 0), n_threads)) {
LLAMA_LOG_ERROR("%s: failed to eval\n", __func__); //LLAMA_LOG_ERROR("%s: failed to decode\n", __func__);
return 1; return 1;
} }
@ -7358,7 +7358,7 @@ int llama_eval_embd(
llama_batch batch = { n_tokens, nullptr, embd, nullptr, nullptr, nullptr, n_past, 1, 0, }; llama_batch batch = { n_tokens, nullptr, embd, nullptr, nullptr, nullptr, n_past, 1, 0, };
if (!llama_decode_internal(*ctx, batch, n_threads)) { if (!llama_decode_internal(*ctx, batch, n_threads)) {
LLAMA_LOG_ERROR("%s: failed to eval\n", __func__); //LLAMA_LOG_ERROR("%s: failed to decode\n", __func__);
return 1; return 1;
} }
@ -7395,7 +7395,7 @@ int llama_decode(
struct llama_batch batch, struct llama_batch batch,
int n_threads) { int n_threads) {
if (!llama_decode_internal(*ctx, batch, n_threads)) { if (!llama_decode_internal(*ctx, batch, n_threads)) {
LLAMA_LOG_ERROR("%s: failed to eval\n", __func__); //LLAMA_LOG_ERROR("%s: failed to decode\n", __func__);
return 1; return 1;
} }