mirror of
https://github.com/LostRuins/koboldcpp.git
synced 2025-09-10 17:14:36 +00:00
Merge commit '13c9a3319b
' into concedo_experimental
# Conflicts: # ggml/src/ggml-cpu/CMakeLists.txt # scripts/sync-ggml.last # tests/test-backend-ops.cpp
This commit is contained in:
commit
ca53d1bedc
18 changed files with 958 additions and 504 deletions
|
@ -218,13 +218,11 @@ struct curl_slist_ptr {
|
|||
#define CURL_MAX_RETRY 3
|
||||
#define CURL_RETRY_DELAY_SECONDS 2
|
||||
|
||||
static bool curl_perform_with_retry(const std::string & url, CURL * curl, int max_attempts, int retry_delay_seconds) {
|
||||
static bool curl_perform_with_retry(const std::string & url, CURL * curl, int max_attempts, int retry_delay_seconds, const char * method_name) {
|
||||
int remaining_attempts = max_attempts;
|
||||
char * method = nullptr;
|
||||
curl_easy_getinfo(curl, CURLINFO_EFFECTIVE_METHOD, &method);
|
||||
|
||||
while (remaining_attempts > 0) {
|
||||
LOG_INF("%s: %s %s (attempt %d of %d)...\n", __func__ , method, url.c_str(), max_attempts - remaining_attempts + 1, max_attempts);
|
||||
LOG_INF("%s: %s %s (attempt %d of %d)...\n", __func__ , method_name, url.c_str(), max_attempts - remaining_attempts + 1, max_attempts);
|
||||
|
||||
CURLcode res = curl_easy_perform(curl);
|
||||
if (res == CURLE_OK) {
|
||||
|
@ -288,13 +286,6 @@ static bool common_download_file_single(const std::string & url, const std::stri
|
|||
try {
|
||||
metadata_in >> metadata;
|
||||
LOG_DBG("%s: previous metadata file found %s: %s\n", __func__, metadata_path.c_str(), metadata.dump().c_str());
|
||||
if (metadata.contains("url") && metadata.at("url").is_string()) {
|
||||
auto previous_url = metadata.at("url").get<std::string>();
|
||||
if (previous_url != url) {
|
||||
LOG_ERR("%s: Model URL mismatch: %s != %s\n", __func__, url.c_str(), previous_url.c_str());
|
||||
return false;
|
||||
}
|
||||
}
|
||||
if (metadata.contains("etag") && metadata.at("etag").is_string()) {
|
||||
etag = metadata.at("etag");
|
||||
}
|
||||
|
@ -302,10 +293,10 @@ static bool common_download_file_single(const std::string & url, const std::stri
|
|||
last_modified = metadata.at("lastModified");
|
||||
}
|
||||
} catch (const nlohmann::json::exception & e) {
|
||||
LOG_ERR("%s: error reading metadata file %s: %s\n", __func__, metadata_path.c_str(), e.what());
|
||||
return false;
|
||||
LOG_ERR("%s: error reading metadata file %s: %s\n", __func__, metadata_path.c_str(), e.what());
|
||||
}
|
||||
}
|
||||
// if we cannot open the metadata file, we assume that the downloaded file is not valid (etag and last-modified are left empty, so we will download it again)
|
||||
} else {
|
||||
LOG_INF("%s: no previous model file found %s\n", __func__, path.c_str());
|
||||
}
|
||||
|
@ -351,7 +342,7 @@ static bool common_download_file_single(const std::string & url, const std::stri
|
|||
|
||||
// we only allow retrying once for HEAD requests
|
||||
// this is for the use case of using running offline (no internet), retrying can be annoying
|
||||
bool was_perform_successful = curl_perform_with_retry(url, curl.get(), 1, 0);
|
||||
bool was_perform_successful = curl_perform_with_retry(url, curl.get(), 1, 0, "HEAD");
|
||||
if (!was_perform_successful) {
|
||||
head_request_ok = false;
|
||||
}
|
||||
|
@ -433,7 +424,7 @@ static bool common_download_file_single(const std::string & url, const std::stri
|
|||
// start the download
|
||||
LOG_INF("%s: trying to download model from %s to %s (server_etag:%s, server_last_modified:%s)...\n", __func__,
|
||||
llama_download_hide_password_in_url(url).c_str(), path.c_str(), headers.etag.c_str(), headers.last_modified.c_str());
|
||||
bool was_perform_successful = curl_perform_with_retry(url, curl.get(), CURL_MAX_RETRY, CURL_RETRY_DELAY_SECONDS);
|
||||
bool was_perform_successful = curl_perform_with_retry(url, curl.get(), CURL_MAX_RETRY, CURL_RETRY_DELAY_SECONDS, "GET");
|
||||
if (!was_perform_successful) {
|
||||
return false;
|
||||
}
|
||||
|
@ -1949,6 +1940,23 @@ common_params_context common_params_parser_init(common_params & params, llama_ex
|
|||
params.sampling.grammar = json_schema_to_grammar(json::parse(value));
|
||||
}
|
||||
).set_sparam());
|
||||
add_opt(common_arg(
|
||||
{"-jf", "--json-schema-file"}, "FILE",
|
||||
"File containing a JSON schema to constrain generations (https://json-schema.org/), e.g. `{}` for any JSON object\nFor schemas w/ external $refs, use --grammar + example/json_schema_to_grammar.py instead",
|
||||
[](common_params & params, const std::string & value) {
|
||||
std::ifstream file(value);
|
||||
if (!file) {
|
||||
throw std::runtime_error(string_format("error: failed to open file '%s'\n", value.c_str()));
|
||||
}
|
||||
std::string schema;
|
||||
std::copy(
|
||||
std::istreambuf_iterator<char>(file),
|
||||
std::istreambuf_iterator<char>(),
|
||||
std::back_inserter(schema)
|
||||
);
|
||||
params.sampling.grammar = json_schema_to_grammar(json::parse(schema));
|
||||
}
|
||||
).set_sparam());
|
||||
add_opt(common_arg(
|
||||
{"--pooling"}, "{none,mean,cls,last,rank}",
|
||||
"pooling type for embeddings, use model default if unspecified",
|
||||
|
|
Loading…
Add table
Add a link
Reference in a new issue