feat: expand model library from 12 to 27 models across 8 families#6
Open
stmatengss wants to merge 6 commits into
Open
feat: expand model library from 12 to 27 models across 8 families#6stmatengss wants to merge 6 commits into
stmatengss wants to merge 6 commits into
Conversation
Add a browser-based KVCache calculator (website/calculator.html) that runs the full estimate-multi-agent heuristic entirely client-side with no backend. Includes preset configs, real-time parameter editing, Chart.js visualization, and TPS capacity planning tables. Also adds a JS parity validation system (tests/validate_js_parity.mjs) that extracts the JS engine from the HTML and validates against Python-generated test vectors to ensure calculation correctness. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Covers: model preset library (10+ models), scenario comparison mode, quick-start wizard, TPS-vs-machines chart, export/share, and UX polish. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Added DeepSeek V3.2, Qwen3.5, Qwen3.6 with verified configs from ModelScope/GitHub. Added MLA and hybrid-attention notes for KV cache sizing. GLM5-5 and DeepSeek V4 marked as TODO (not yet public). Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
…rt and comparison Add 6 major features to the interactive web calculator: - Model Preset Library with 12+ models (Qwen, Llama, DeepSeek, Mistral, GLM, Yi) - Quick-Start Wizard with 3-step guided flow (model, hardware, workload) - TPS vs Machine Count chart and parameter sensitivity analysis - Export & Share (Markdown, PNG, JSON, URL hash with lz-string compression) - Scenario Comparison mode with delta highlighting - UX polish: dark/light mode, input validation, tooltips, responsive layout Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Expand GPU support beyond H20-only configurations: - Enrich GPU_SPECS with 9 GPU types (H20, H100, H100 NVL, H200, A100-80G, A100-40G, A800, L40S, L20) including topology and baseline TPS metadata - Add GPU hardware dropdown to Deployment form for quick auto-fill - Add 5 model×hardware deployment presets (Llama 70B/A100, Qwen2.5-72B/H100, Mistral 7B/L40S, Llama 8B/A100, DeepSeek-V3/H200) - Update wizard to use per-GPU cards_per_machine and baseline_tps_per_card Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Add 15 new model entries sourced from HuggingFace config.json: - Qwen3: 235B-A22B (MoE), 32B, 14B, 8B, 4B, 1.7B - Llama 4: Scout (17Bx16E), Maverick (17Bx128E) - DeepSeek: R1 685B - Mistral: Small 24B, Nemo 12B - Phi: Phi-4 14B, Phi-3 Medium 14B - InternLM: 2.5-20B, 2.5-7B Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
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.
Summary
config.json:Test plan
🤖 Generated with Claude Code