Commit Graph

Select branches

Hide Pull Requests

AUTO-664--Healthcheck-error

AUTO-695

AUTO-702--improved-test-load

AUTO-703-comfyui-json-error

AUTO-848--endpoint-id-routing

AUTO-874--fix-openai-worker-client

AUTO-912

AUTO-comfyui-updates

AUTO-tgi-client-edits

HOTFIX-aiohttp-version

add-hacky-deployments-script

add-llama-log

add-ntlk-requirement

allow-prerelease-pip

batched-wait-time

blindhotfix_comfy_ui_default_port

bugfix/backend-timeout-infinite

bugfix/comfyui-json-cost-fix

bugfix/startup-script-20250901

comfy-blind-fix-override

comfy-queue

comfy-request-idx

comfyui-benchmark-model-type

delete-reqs-hotfix

deployments-direct-executable

experiment/null-worker

feat/comfyui-api-s3-webhook

fifo-queue

fix-bad-model-rotate

fix-signature

fix-vllm-concurrency

fix/better-benchmarks

fix/comfyui-restore-benchmark-json-loading

fix/fetch-vast-certificate-for-pyworker-client

fix/fix-pyright-and-revert-cancelled-http-request-handling

fix/more-robust-tracking-of-pending-request

fix/pyworker-num-requests-working

fix/reduce-dependencies

force-update-flag

hotfix-pin-pycares

hotfix-vllm-parsing

kill-pub-sub

main

model-log-rotation

new-pyworker

new-pyworker-vllm-cold-start-timing

new-pyworker-vllm-prefix-cache

no-input

obfuscate-mtoken

pyworker-failure-reporting

pyworker-force-update-and-retry

pyworker-queue

pyworker-sdk

read-model-log

remove-sleeps-and-delays

script-testing-new-pyworker

sdk-versioning

send-metrics-on-req-start

serverside-sdk

session

shallow-git-clone

synthesis

synthesis_fix

test-package

unsecure_report_addr

update-client-scripts

update-requirements

use-mtoken

use-vastai

Mono Color

• ac1e109c48 Merge pull request #47 from vast-ai/new-pyworker-vllm-prefix-cache Colter-Downing 2025-10-27 12:30:34 -07:00
• d6eb498ee4 catch the case where all benchmarks fail (sets error) new-pyworker-vllm-prefix-cache Colter Downing 2025-10-27 12:01:55 -07:00
• f56bbc0ebe Added request_idx to comfy auth_data Abiola Akinnubi 2025-10-27 03:17:06 +00:00
• 5d5bc197d7 adding timings for cold start Colter Downing 2025-10-26 18:44:23 -07:00
• bcecd6df40 Suppress matplot debug logs Colter Downing 2025-10-25 16:18:02 -07:00
• e756f61b9a graphing errors over time script-testing-new-pyworker Colter Downing 2025-10-25 12:14:27 -07:00
• 8cb98c84f9 non vibe coded test_load Colter Downing 2025-10-24 19:08:36 -07:00
• 4d9bf2048c Fix Lucas Armand 2025-10-24 15:44:38 -07:00
• 7788bc4a62 Added some debug logs Lucas Armand 2025-10-24 15:41:00 -07:00
• e251afda2b improved test load Colter Downing 2025-10-09 19:37:39 -07:00
• 74bd932327 Suppress matplot debug logs Lucas Armand 2025-10-10 11:57:46 -07:00
• 37ad3f8d46 asyncio in metrics Lucas Armand 2025-10-23 10:18:31 -07:00
• 70d51bafe1 Merge pull request #36 from robballantyne/feat/comfyui-json-benchmark-workflow-from-file Rob Ballantyne 2025-10-23 17:05:48 +01:00
• 63909736bb Merge pull request #4 from robballantyne/feat/comfyui-json-benchmark-workflow-from-file-no-silent-fail Rob Ballantyne 2025-10-23 17:02:12 +01:00
• f4f7080df1 Re-add comment Rob Ballantyne 2025-10-23 17:00:28 +01:00
• d51a338e8f log when benchmark file not used Rob Ballantyne 2025-10-23 16:41:02 +01:00
• 92a04bd7af No silent fail if benchmark file is missing Rob Ballantyne 2025-10-23 13:41:03 +01:00
• 0f13506938 Send success param Lucas Armand 2025-10-22 10:18:59 -07:00
• 01e752d31f use more asyncio sleep Lucas Armand 2025-10-21 18:52:13 -07:00
• 5edfa968ca async sleep Lucas Armand 2025-10-21 18:49:48 -07:00
• 5b5ef7227a nvm moved it here Lucas Armand 2025-10-21 18:20:11 -07:00
• 16990ff8ff move start request Lucas Armand 2025-10-21 18:18:44 -07:00
• 9748176366 fixed semaphore acquire bool Lucas Armand 2025-10-21 18:12:23 -07:00
• b39193ae70 check for sem acquire Lucas Armand 2025-10-21 18:02:14 -07:00
• 9a6ca5d412 added versioning Lucas Armand 2025-10-21 15:42:43 -07:00
• e9ba1b03e4 Use delete_requests and track request_idxs Lucas Armand 2025-10-08 16:54:18 -07:00
• a7617162a7 spelling fix serverside-sdk Lucas Armand 2025-10-15 15:09:34 -07:00
• d8f51a2edc updated startserver Lucas Armand 2025-10-15 12:14:31 -07:00
• ee57ed207b stat script Lucas Armand 2025-10-14 17:38:06 -07:00
• 44bfe634f5 do a shallow git clone of the pyworker shallow-git-clone Lucas Armand 2025-10-14 13:51:04 -07:00
• c98d661513 Merge pull request #39 from vast-ai/remove-time-divide LucasArmandVast 2025-10-13 10:06:22 -07:00
• 3988cf553f Suppress matplot debug logs AUTO-702--improved-test-load Lucas Armand 2025-10-10 11:57:46 -07:00
• a00c1adab5 improved test load Colter Downing 2025-10-09 19:37:39 -07:00
• f6fd1c6ac1 merge Lucas Armand 2025-10-09 18:15:55 -07:00
• 055e346c8c Send metrics on request start send-metrics-on-req-start Lucas Armand 2025-10-09 10:13:50 -07:00
• 1cedb28acf Removed division by elapsed time, since autoscaler cur_load in units of workload Lucas Armand 2025-10-08 16:54:18 -07:00
• 7d3be849d9 Handle errors from model for comfyui-json AUTO-703-comfyui-json-error Lucas Armand 2025-10-08 12:00:45 -07:00
• ec25dda3ad Merge branch 'vast-ai:main' into feat/comfyui-json-benchmark-workflow-from-file Rob Ballantyne 2025-10-08 14:49:32 +01:00
• 0397af719d Merge pull request #37 from robballantyne/bugfix/healthcheck-endpoint Colter-Downing 2025-10-06 15:11:27 -07:00
• 4fdc314fd9 Fix healthcheck endpoint URL Rob Ballantyne 2025-10-06 22:16:09 +01:00
• 3786cf978d Add awareness of errors thrown by the provisioning script Rob Ballantyne 2025-10-05 23:14:59 +01:00
• a86d4bcf9c Import json Rob Ballantyne 2025-10-05 23:05:33 +01:00
• e9b6a14a5e Import Path Rob Ballantyne 2025-10-05 22:59:19 +01:00
• cadac033e1 Enables use of custom workflow for benchmarking Retains existing method is misc/benchmark.json is nopt present Rob Ballantyne 2025-10-05 22:53:22 +01:00
• 639d82f5b4 Merge pull request #35 from vast-ai/AUTO-664--Healthcheck-error Colter-Downing 2025-10-02 12:51:19 -07:00
• 25db78e39d Fix healthcheck with separate session AUTO-664--Healthcheck-error Colter Downing 2025-10-01 18:04:31 -07:00
• 4e2f2311d0 Merge pull request #33 from vast-ai/comfy-blind-fix-override Scott-Laytart 2025-09-03 11:50:07 -07:00
• 38782d89bc undo the fix for comfy yesterday. comfy-blind-fix-override abiola-vastai 2025-09-03 17:12:35 +00:00
• 0185216ccb Merge pull request #32 from vast-ai/blindhotfix_comfy_ui_default_port Scott-Laytart 2025-09-02 18:26:25 -07:00
• b20d9e714c Blind hotfix to see if comfy UI default is needed. if it does work we would revert back. blindhotfix_comfy_ui_default_port abiola-vastai 2025-09-03 01:20:09 +00:00
• b1eb65d75d Merge pull request #31 from vast-ai/bugfix/startup-script-20250901 Rob Ballantyne 2025-09-01 18:19:17 +01:00
• 1d09d7fe96 Update uv venv creation command bugfix/startup-script-20250901 Rob Ballantyne 2025-09-01 16:55:20 +01:00
• 1b37054dec Merge pull request #28 from vast-ai/bugfix/backend-timeout-infinite Colter-Downing 2025-08-28 11:22:33 -07:00
• 1a1e4174b8 Merge pull request #29 from vast-ai/bugfix/comfyui-json-cost-fix Colter-Downing 2025-08-28 11:22:21 -07:00
• b8377c4081 Set cost to 100 bugfix/comfyui-json-cost-fix Rob Ballantyne 2025-08-28 16:13:17 +01:00
• 1e4fa87437 Prevent timeout and allow long running connections bugfix/backend-timeout-infinite Rob Ballantyne 2025-08-28 15:48:57 +01:00
• 4c5fa03c7b adds import for ClientTimeout Rob Ballantyne 2025-08-27 20:54:27 +01:00
• a8fe74f771 Remove default 300s timeout Rob Ballantyne 2025-08-27 18:34:45 +01:00
• b482de8394 Merge pull request #27 from vast-ai/feat/comfyui-api-s3-webhook Rob Ballantyne 2025-08-26 14:22:05 +01:00
• 703435d10e Improve MODEL_SERVER_START_* messages feat/comfyui-api-s3-webhook Rob Ballantyne 2025-08-26 12:42:04 +01:00
• 947fc5eea4 Improve benchmarking explanation Rob Ballantyne 2025-08-26 12:41:30 +01:00
• 7c1a544b19 Improve error reporting when no ready workers Rob Ballantyne 2025-08-26 12:41:05 +01:00
• 16b414676e Use count_workload() function for cost Rob Ballantyne 2025-08-25 18:31:10 +01:00
• ba74ac8136 Use cost value 1 for all jobs Rob Ballantyne 2025-08-25 17:58:22 +01:00
• 92ff412679 Use MODEL_SERVER_URL environment variable Rob Ballantyne 2025-08-25 17:57:32 +01:00
• fc75a64684 Use MODEL_SERVER_URL environment variable Rob Ballantyne 2025-08-25 17:56:27 +01:00
• b00bef547c Ensure uv env script is present before sourcing Rob Ballantyne 2025-08-22 17:08:42 +01:00
• 3f4acb29fa Improved client exception handling Rob Ballantyne 2025-08-22 15:20:15 +01:00
• 58b078f908 Fix modifier class Rob Ballantyne 2025-08-20 18:06:02 +01:00
• f9fdf04884 Fix signature Rob Ballantyne 2025-08-20 13:27:29 +01:00
• 636f17d27f Fix workflow modifier class Rob Ballantyne 2025-08-20 09:57:07 +01:00
• 08c88f7527 Improve testability Rob Ballantyne 2025-08-20 09:34:09 +01:00
• 8797b504af Initial ComfyUI implementation with updated wrapper Rob Ballantyne 2025-08-19 17:59:20 +01:00
• cd946b0a9f update report_addr to use new webserver endpoint with AS fallback Nader Arbabian 2025-08-12 13:31:19 -07:00
• c595b42410 for benchmarking, use concurrent requests (#26) Nader Arbabian 2025-08-11 12:39:28 -07:00
• 0bf3247a34 fix completions and interactive client Nader Arbabian 2025-08-11 12:37:53 -07:00
• eecefd1d52 for benchmarking, use concurrent requests fix/better-benchmarks Nader Arbabian 2025-08-11 12:14:32 -07:00
• 52ac4c0c1a fix endpoint_util not using the correct instance's endpoint Nader Arbabian 2025-08-11 11:49:51 -07:00
• 8804e17201 download vast.ai's root certificate in order to make pyworker requests (#25) Nader Arbabian 2025-08-08 17:04:16 -07:00
• 4016cf9a53 redo metrics tracking for requests, fixes bug wherere some requests were marked as pending, even though they had finished (#24) Nader Arbabian 2025-08-08 17:01:21 -07:00
• 9773e5f67b download vast.ai's root certificate in order to make pyworker requests fix/fetch-vast-certificate-for-pyworker-client Nader Arbabian 2025-07-31 12:47:12 -07:00
• d3be9fe7db redo metrics tracking for requests, fixes bug wherere some requests were marked as pending, even though they had finished fix/more-robust-tracking-of-pending-request Nader Arbabian 2025-07-30 18:56:51 -07:00
• e0be45f39a Addresses breaking change in core pyworker (#22) Rob Ballantyne 2025-07-18 01:09:23 +01:00
• be2aafdb1f fix pyright errors + revert to old way of handling cancelled api requests (#23) Nader Arbabian 2025-07-17 16:59:06 -07:00
• 4ac51947b4 fix pyright errors + revert to old way of handling cancelled api requests fix/fix-pyright-and-revert-cancelled-http-request-handling Nader Arbabian 2025-07-17 15:18:21 -07:00
• 9e369c55a5 Ensure venv creation where python is unavailable (#21) Rob Ballantyne 2025-07-17 17:59:35 +01:00
• 69d9b7455f OpenAI compatible worker (#19) Rob Ballantyne 2025-07-16 09:46:26 +01:00
• 6fb610cb5b fix pyworker miscounting active connections (#20) Nader Arbabian 2025-07-15 15:33:27 -07:00
• 6b0f019cf7 add option to skip auth fix/pyworker-num-requests-working Nader Arbabian 2025-07-15 14:02:33 -07:00
• ce52419023 AUTO-421: clean up some issues Nader Arbabian 2025-07-11 15:04:54 -07:00
• 3e49b7d04b AUTO-421: fix pyworker miscounting active connections Nader Arbabian 2025-07-10 19:27:55 -07:00
• 0bf2d04223 stop using urljoin for worker_status endpoint Nader Arbabian 2025-06-17 23:09:45 -07:00
• 9ebf1924ea don't healthcheck endpoints until model is loaded and benchmarks have run Nader Arbabian 2025-06-11 15:25:57 -07:00
• 0ab9a13a46 update tokenizers deps Nader Arbabian 2025-06-10 17:53:01 -07:00
• 72a5f6ad13 update tokenizers deps fix/reduce-dependencies Nader Arbabian 2025-06-10 17:53:01 -07:00
• 4bac805093 update tokenizers Nader Arbabian 2025-06-10 17:01:28 -07:00
• d99adcfb36 build(deps): bump transformers from 4.44.2 to 4.50.0 dependabot[bot] 2025-06-10 22:08:37 +00:00
• 4bf6f268a2 fix up depencies once and for all, fix broken imports Nader Arbabian 2025-06-10 14:09:38 -07:00
• 90877b758b build(deps): bump requests from 2.32.3 to 2.32.4 dependabot[bot] 2025-06-09 22:02:27 +00:00
• f7cfcb0a66 build(deps): bump setuptools from 70.3.0 to 78.1.1 dependabot[bot] 2025-06-09 21:50:04 +00:00