docs: refresh README onboarding + fix CLI doc inconsistencies#485
Open
iamagenius00 wants to merge 1 commit into
Open
docs: refresh README onboarding + fix CLI doc inconsistencies#485iamagenius00 wants to merge 1 commit into
iamagenius00 wants to merge 1 commit into
Conversation
This is a documentation-only change. No source code or behavior is
affected. Three groups of edits:
## 1. Fix CLI / version inconsistencies found against source code
- `docs/user_guide/work_with_openclaw.md`: remove `--enable-prefix-cache`
from both `parallax join` examples. The flag does not exist in
`src/parallax/server/server_args.py`; only `--disable-prefix-cache`
is defined (prefix cache is enabled by default). The OpenClaw guide
commands currently fail with `argparse: unrecognized arguments` when
copy-pasted verbatim.
- `docs/user_guide/work_with_openclaw.md`: bump the Python prerequisite
to `>=3.11,<3.14` to match `pyproject.toml` and `install.md`.
- `docs/user_guide/quick_start.md`: fix `parallax serve --m` to
`--model-path`. `--m` only worked through argparse abbreviation
matching and silently breaks if any new `--m*` flag is added.
- `docs/user_guide/install.md`: clarify that the Ubuntu-24.04
requirement applies to Blackwell GPUs (B100/B200/RTX 50) from source,
not to every install path; cross-reference the Docker section.
- `docs/user_guide/install.md`: add a note that the bundled `vllm-rs`
Rust HTTP frontend is built from vLLM source at `v0.22.0` (set by
`VLLM_REF` in `install.sh`), and that this is independent of the
optional `vllm` Python extras (`vllm==0.14.0`) used for the GPU
backend. The version mismatch confuses readers today.
## 2. Restructure README onboarding
- Replace the separate `User Guide` and `Quick Install` sections with
a single `Getting Started` flow:
- Step 0: install path table + the auto-detecting `./install.sh`
command, with a link to the full Installation Guide for Windows,
Docker, and Python version selection.
- Step 1: `parallax run` on the scheduler machine.
- Step 2: open the dashboard, choose a model, with a screenshot.
- Step 3: `parallax join` on each worker, with a screenshot.
- Step 4: chat or call the local OpenAI-compatible endpoint at
`http://localhost:3001/v1`, with the chat-interface screenshot.
- A trailing link points to the full quick_start guide for remote
mode, no-frontend setup, standalone serving, and API examples.
- Move FAQ from `quick_start.md` into the README at the end of the
file, wrapped in `<details>` collapsible blocks so the README stays
short by default. The README is now the single source of truth for
FAQ; the duplicate section in `quick_start.md` is removed.
- News: add the ModelScope downloads entry. Drop one earlier draft
entry that was not backed by code.
## 3. Use absolute image URLs in README
Convert every `docs/images/...` reference in the README to
`https://raw.githubusercontent.com/GradientHQ/parallax/main/docs/images/...`,
including the hero image, the seven partner logos, and the three
Getting Started screenshots. Relative paths render on github.com but
break on PyPI and other downstream README renderers.
## Supported Models
Add a row for StepFun / Step-3.5-Flash. The model is already
registered in `src/backend/server/static_config.py` and has a
dedicated architecture file at `src/parallax/models/step3p5.py`, but
was missing from the README table.
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
1 participant
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.
Two things bundled:
CLI / version doc fixes found while reading source vs docs:
work_with_openclaw.md:--enable-prefix-cachedoesn'''t exist (prefix cache is on by default; only--disable-prefix-cacheis defined). Removed from both examples.quick_start.md:parallax serve --m→--model-path.work_with_openclaw.md: Python prereq →>=3.11,<3.14to matchpyproject.toml.install.md: clarify Ubuntu-24.04 is for Blackwell GPUs from source; add note that bundledvllm-rs(v0.22.0) ≠ optionalvllmPython extras (0.14.0).README onboarding refresh:
Quick Install+User Guideinto a singleGetting Startedflow (Step 0–4) with the dashboard / join / chat screenshots embedded inline.quick_start.mdinto collapsible<details>blocks at the end of the README so it'''s discoverable without bloating the body.static_config.py).docs/images/...references to absoluteraw.githubusercontent.comURLs so the README renders on PyPI too.No code changes.