Repository navigation
docs: sync model catalog with dashboard API - #64
Merged
Merged
Conversation
- gliner2.5-decide-v1: input $0.004 -> $0.03 (catalog table + card pill, text-classification table) - add d1.1 (Text Classification, decision, 65,536 max tokens, $0.03 input, output not billed): catalog row + card + task-library count 8 -> 9, text-classification table row and `## d1.1` section, model page, d1_1 playground spec, SystemOneRequest.model, docs.json nav, and the System One reference page (four models -> five) - audit script: check SystemOneRequest.model for decision models and the audio request enums for transcription/speech models, instead of the Responses/Chat enums they are not routable on d1.1 is routable only on POST /v1/systemone per the API description, and the payload ships no sample bodies for it, so its section documents the request only and no response example was invented. No models were removed; the API returns 31. Source: https://api-dashboard.zerogpu.ai/api/models Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01V6BqNyrYRBCR1fUq6WhACe
Contributor
|
Preview deployment for your docs. Learn more about Mintlify Previews.
💡 Tip: Enable Automations to automatically generate PRs for you. |
|
Important
This repository does not receive automatic reviews because it has fewer than 10 stars.
|
This branch was successfully deployed
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Automated model-catalog sync. The dashboard API is the source of truth; every value below was taken from it. The API returned 31 models.
Corrected
gliner2.5-decide-v1Fixed in the
docs/model-catalog.mdxtable cell and its card pill, and in thedocs/text-classification.mdxtable cell. No other page quoted the old price.Added
d1.1— Text Classification (taskType: decision), 65,536 max tokens, $0.03 / 1M input, output not billed, no cached-input rate. All seven surfaces:docs/model-catalog.mdx— At-a-glance row, detailed card, and the Text Classification library card (8 → 9 models). A note below the table explains thatd1.1is called on/v1/systemoneand bills input only.docs/text-classification.mdx— table row and a## d1.1section; the intro now carves outd1.1from the "Responses + Chat Completions" claim.api-reference/models/d1-1.mdx— new page, pinned toPOST /systemone.api-reference/openapi/playgrounds/d1_1.openapi.json— new spec, cloned fromzlm-v1-moderation-edge.api-reference/openapi/zerogpu.openapi.json—d1.1added toSystemOneRequest.model.docs.json— nav entry under API Reference → By model, beside the other decision models.api-reference/systemone.mdx— table row and "four models" → "five models".Removed
None — the API still returns every documented model.
Judgment calls
sample_responses_bodyand nosample_chat_completions_bodyford1.1, and its own description states it is called onPOST /v1/systemoneand "does not chat, stream or call tools". It was therefore wired to/systemoneonly, not added to the Responses or Chat enums. The skill's "chat-only when there is no Responses sample" rule assumes a chat sample exists; it does not here.## d1.1section documents the request and what comes back in prose, with no JSON response example, and the playground's 200 shows the schema without a fabricated example payload./systemoneschema already in this repo, using the use cases the API description names (intent, routing, triage).[Model docs]omitted.model_doc_urlpoints back atdocs.zerogpu.aiandterms_url(www.liquid.ai/terms-conditions) carries no model-card path, so per the skill the link is dropped and Terms + Privacy are kept.d1.1hasdisplayPriority: 30, the highest in the catalog, so it leads both tables; no existing row was reshuffled.zlm-v1-moderation-edge; the/moderationspath and the seven schemas only it referenced were dropped, sinced1.1is not routable there. Every System One and error schema is byte-identical to the sibling, and the spec has no dangling$ref.audit-models.pyonly knew the Responses/Chat/moderation/embedding enums, so it reportedd1.1,whisper-tiny,chatterbox-nanoandchatterboxas absent from "both enums" although all four are correctly wired to their own endpoints. It now checksSystemOneRequest.modelfor decision models and the transcription/speech enums for audio models, in the same shape as the existing moderation and embedding cases. Those three audio findings predate this run.Verification
audit-models.py --strict— 0 findings, clean apart from threeNOTElines for the unmappedSpeech-to-Text/Text-to-Speechtasks, whose task page (docs/audio.mdx) already exists and lists all three models.docs.json,zerogpu.openapi.jsonand every playground spec parse./api-reference/models/<slug>link resolves to a file; the seven that do not areredirectssources from earlier removals.N models:count matches the ids beside it (10, 16, 3, 9, 1, 2, 4).<Card href>and nav entry resolves.Generated by Claude Code