vbatts/llama.cpp
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

ci: bench: fix finish reason rate

Browse source
This commit is contained in:
Pierrick HYMBERT 2024-04-05 01:30:44 +02:00
parent b6b50b11f9
commit 8789e17f8f
1 changed files with 6 additions and 2 deletions
Download patch file Download diff file Expand all files Collapse all files

8
examples/server/bench/script.js
View file

@ -99,6 +99,7 @@ export default function () {
let promptEvalEndTime = null let promptEvalEndTime = null
let prompt_tokens = 0 let prompt_tokens = 0
let completions_tokens = 0 let completions_tokens = 0
let finish_reason = null
const res = sse.open(`${server_url}/chat/completions`, params, function (client) { const res = sse.open(`${server_url}/chat/completions`, params, function (client) {
client.on('event', function (event) { client.on('event', function (event) {
if (promptEvalEndTime == null) { if (promptEvalEndTime == null) {
@ -107,8 +108,9 @@ export default function () {
let chunk = JSON.parse(event.data) let chunk = JSON.parse(event.data)
let choice = chunk.choices[0] let choice = chunk.choices[0]
llamacpp_completions_truncated_rate.add(choice.finish_reason === 'length') if (choice.finish_reason) {
llamacpp_completions_stop_rate.add(choice.finish_reason === 'stop') finish_reason = choice.finish_reason
}
if (chunk.usage) { if (chunk.usage) {
prompt_tokens = chunk.usage.prompt_tokens prompt_tokens = chunk.usage.prompt_tokens
@ -140,6 +142,8 @@ export default function () {
if (completions_tokens > 0 && completion_time > 0) { if (completions_tokens > 0 && completion_time > 0) {
llamacpp_tokens_second.add(completions_tokens / completion_time * 1.e3) llamacpp_tokens_second.add(completions_tokens / completion_time * 1.e3)
} }
llamacpp_completions_truncated_rate.add(finish_reason === 'length')
llamacpp_completions_stop_rate.add(finish_reason === 'stop')
sleep(0.3) sleep(0.3)
} }