huggingface/text-generation-inference
5
0
Fork 0
mirror of https://github.com/huggingface/text-generation-inference.git synced 2025-09-12 12:54:52 +00:00
Code Issues Packages Projects Releases Wiki Activity

only tokenizer class instead of hole struct

Browse Source
This commit is contained in:
ErikKaumk 2024-07-11 15:40:30 +02:00
parent f965deeee7
commit 30d86ed95a
3 changed files with 5 additions and 12 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

10
docs/source/usage_statistics.md
View File

@ -33,15 +33,7 @@ As of release 2.1.2 this is an example of the data collected:
},
"ngrok": false,
"revision": null,
"tokenizer_config": {
"add_bos_token": null,
"add_eos_token": null,
"bos_token": "<s>",
"chat_template": null,
"completion_template": null,
"eos_token": "</s>",
"tokenizer_class": "BloomTokenizerFast"
},
"tokenizer_class": "BloomTokenizerFast",
"validation_workers": 2,
"waiting_served_ratio": 1.2,
"docker_label": "latest",

3
router/src/main.rs
View File

@ -327,6 +327,7 @@ async fn main() -> Result<(), RouterError> {
tracing::warn!("Could not find tokenizer config locally and no API specified");
HubTokenizerConfig::default()
});
let tokenizer_class = tokenizer_config.tokenizer_class.clone();
let tokenizer: Option<Tokenizer> = tokenizer_filename.and_then(|filename| {
let mut tokenizer = Tokenizer::from_file(filename).ok();
@ -387,7 +388,7 @@ async fn main() -> Result<(), RouterError> {
let user_agent = if !disable_usage_stats && is_docker {
let reducded_args = usage_stats::Args::new(
config.clone(),
tokenizer_config.clone(),
tokenizer_class,
max_concurrent_requests,
max_best_of,
max_stop_sequences,

4
router/src/usage_stats.rs
View File

@ -60,7 +60,7 @@ impl UsageStatsEvent {
#[derive(Debug, Clone, Serialize)]
pub struct Args {
model_config: Option<Config>,
tokenizer_config: HubTokenizerConfig,
tokenizer_config: Option<String>,
max_concurrent_requests: usize,
max_best_of: usize,
max_stop_sequences: usize,
@ -86,7 +86,7 @@ pub struct Args {
impl Args {
pub fn new(
model_config: Option<Config>,
tokenizer_config: HubTokenizerConfig,
tokenizer_config: Option<String>,
max_concurrent_requests: usize,
max_best_of: usize,
max_stop_sequences: usize,