From 0577778fda95dd8849bbf1bb2be07f61125b992d Mon Sep 17 00:00:00 2001 From: Jun Siang Cheah Date: Tue, 5 Nov 2024 23:29:01 +0000 Subject: [PATCH] docs: update models --- .github/workflows/update-readme.yml | 1 + README.md | 39 ++++++++------ src/pull_available_models.py | 82 ++++++++++++++++++++++++----- 3 files changed, 94 insertions(+), 28 deletions(-) diff --git a/.github/workflows/update-readme.yml b/.github/workflows/update-readme.yml index dca6e14..e8570fc 100644 --- a/.github/workflows/update-readme.yml +++ b/.github/workflows/update-readme.yml @@ -35,6 +35,7 @@ jobs: HYPERBOLIC_API_KEY: ${{ secrets.HYPERBOLIC_API_KEY }} GCP_PROJECT_ID: ${{ secrets.GCP_PROJECT_ID }} GOOGLE_APPLICATION_CREDENTIALS: ${{ steps.auth.outputs.credentials_file_path }} + LAMBDA_API_KEY: ${{ secrets.LAMBDA_API_KEY }} run: | python -u src/pull_available_models.py - name: Remove credentials diff --git a/README.md b/README.md index d637706..1cea985 100644 --- a/README.md +++ b/README.md @@ -50,6 +50,7 @@ This lists various services that provide free access or credits towards API-base Llama 3.2 11B Vision Instruct Llama 3.2 1B Instruct Llama 3.2 3B Instruct +Llama 3.2 90B Vision Instruct Mistral 7B Instruct Mythomax L2 13B Mythomist 7B @@ -59,7 +60,6 @@ This lists various services that provide free access or credits towards API-base Qwen 2 7B Instruct Toppy M 7B Zephyr 7B Beta -meta-llama/llama-3.2-90b-vision-instruct:free Google AI Studio Data is used for training (when used outside of the UK/CH/EEA/EU). @@ -96,17 +96,18 @@ This lists various services that provide free access or credits towards API-base embedding-001 - - Lambda Labs (Free Preview) - Free for a limited time - Nous Hermes 3 Llama 3.1 405B (FP8) - - - - - Liquid LFM 40B - - + Lambda Labs (Free Preview)Requires credit card verification.DeepSeek Coder v2 Lite Instruct +Dracarys 2 72B Instruct +Hermes 3 405B +Hermes 3 70B +Hermes 3 8B +Liquid LFM 40B +Llama 3.1 405B Instruct (FP8) +Llama 3.1 70B Instruct (FP8) +Llama 3.1 8B Instruct +Llama 3.1 Nemotron 70B Instruct +Llama 3.2 3B Instruct + Mistral (La Plateforme) Free tier (Experiment plan) requires opting into data training, requires phone number verification. Open and Proprietary Mistral models @@ -122,15 +123,23 @@ This lists various services that provide free access or credits towards API-base Various open models 1,000 requests/day (with an account) - SambaNova Cloud - + SambaNova Cloud + Llama 3.1 405B 10 requests/minute + + Llama 3.2 90B + 1 request/minute + Llama 3.1 70B 20 requests/minute + + Llama 3.2 11B + 10 requests/minute + Llama 3.1 8B 30 requests/minute @@ -177,6 +186,7 @@ This lists various services that provide free access or credits towards API-base Phi-3-small instruct (8k) Phi-3.5-mini instruct (128k) OVH AI Endpoints (Free Alpha)Token expires every week.CodeLlama 13B Instruct12 requests/minute +Codestral Mamba 7B v0.112 requests/minute Llama 2 13B Chat12 requests/minute Llama 3 70B Instruct12 requests/minute Llama 3 8B Instruct12 requests/minute @@ -185,7 +195,6 @@ This lists various services that provide free access or credits towards API-base Mistral 7B Instruct12 requests/minute Mixtral 8x22B Instruct12 requests/minute Mixtral 8x7B Instruct12 requests/minute -mamba-codestral-7b-v0-112 requests/minute Cloudflare Workers AI10,000 tokens/dayDeepseek Coder 6.7B Base (AWQ) Deepseek Coder 6.7B Instruct (AWQ) Deepseek Math 7B Instruct diff --git a/src/pull_available_models.py b/src/pull_available_models.py index 398c553..e7ea949 100644 --- a/src/pull_available_models.py +++ b/src/pull_available_models.py @@ -128,6 +128,20 @@ MODEL_TO_NAME_MAPPING = { "mathstral-7b-v0.1": "Mathstral 7B v0.1", "llama-3.1-70b-instruct": "Llama 3.1 70B Instruct", "gryphe/mythomax-l2-13b:free": "Mythomax L2 13B", + "meta-llama/llama-3.2-90b-vision-instruct:free": "Llama 3.2 90B Vision Instruct", + "mamba-codestral-7b-v0-1": "Codestral Mamba 7B v0.1", + "hermes3-70b": "Hermes 3 70B", + "llama3.1-nemotron-70b-instruct": "Llama 3.1 Nemotron 70B Instruct", + "llama3.2-3b-instruct": "Llama 3.2 3B Instruct", + "llama3.1-8b-instruct": "Llama 3.1 8B Instruct", + "llama3.1-70b-instruct-fp8": "Llama 3.1 70B Instruct (FP8)", + "llama3.1-405b-instruct-fp8": "Llama 3.1 405B Instruct (FP8)", + "hermes3-405b": "Hermes 3 405B", + "deepseek-coder-v2-lite-instruct": "DeepSeek Coder v2 Lite Instruct", + "hermes3-8b": "Hermes 3 8B", + "dracarys2-72b-instruct": "Dracarys 2 72B Instruct", + "lfm-40b": "Liquid LFM 40B", + } @@ -151,6 +165,11 @@ HYPERBOLIC_IGNORED_MODELS = { "TTS" } +LAMBDA_IGNORED_MODELS = { + "lfm-40b-vllm", + "hermes3-405b-fp8-128k" +} + def get_model_name(id): id = id.lower() @@ -496,6 +515,32 @@ def fetch_gemini_limits(logger): return models +def fetch_lambda_models(logger): + logger.info("Fetching Lambda Labs models...") + r = requests.get( + "https://api.lambdalabs.com/v1/models", + headers={ + "Authorization": f"Bearer {os.environ['LAMBDA_API_KEY']}", + }, + ) + r.raise_for_status() + models = r.json()["data"] + logger.info(f"Fetched {len(models)} models from Lambda Labs") + ret_models = [] + for model in models: + if model["id"] in LAMBDA_IGNORED_MODELS: + logger.debug(f"Ignoring model {model['id']}") + continue + ret_models.append( + { + "id": model["id"], + "name": get_model_name(model["id"]), + } + ) + ret_models = sorted(ret_models, key=lambda x: x["name"]) + return ret_models + + def get_human_limits(model): if "limits" not in model: return "" @@ -512,6 +557,7 @@ def main(): cloudflare_logger = create_logger("Cloudflare") github_logger = create_logger("GitHub") hyperbolic_logger = create_logger("Hyperbolic") + lambda_logger = create_logger("Lambda Labs") gemini_models = fetch_gemini_limits(google_ai_studio_logger) openrouter_models = fetch_openrouter_models(openrouter_logger) @@ -519,6 +565,7 @@ def main(): ovh_models = fetch_ovh_models(ovh_logger) cloudflare_models = fetch_cloudflare_models(cloudflare_logger) github_models = fetch_github_models(github_logger) + lambda_models = fetch_lambda_models(lambda_logger) groq_models = fetch_groq_models(groq_logger) @@ -598,17 +645,18 @@ def main(): embedding-001 """ - table += """ - Lambda Labs (Free Preview) - Free for a limited time - Nous Hermes 3 Llama 3.1 405B (FP8) - - - - - Liquid LFM 40B - - """ + for idx, model in enumerate(lambda_models): + table += "" + if idx == 0: + table += f'' + table += 'Lambda Labs (Free Preview)' + table += '' + table += f'' + table += 'Requires credit card verification.' + table += '' + table += f"{model['name']}" + table += "" + table += "\n" table += """ Mistral (La Plateforme) @@ -632,15 +680,23 @@ def main(): """ table += """ - SambaNova Cloud - + SambaNova Cloud + Llama 3.1 405B 10 requests/minute + + Llama 3.2 90B + 1 request/minute + Llama 3.1 70B 20 requests/minute + + Llama 3.2 11B + 10 requests/minute + Llama 3.1 8B 30 requests/minute