From 5ca325c7257b71fc4cd12ee26bde3e992ade5756 Mon Sep 17 00:00:00 2001 From: Marco Vinciguerra Date: Fri, 10 Jan 2025 09:27:57 +0100 Subject: [PATCH] docs: refactoring of the doc --- docs/source/introduction/overview.rst | 79 ++++++++++++++++--- .../scrapegraphai.helpers.models_tokens.rst | 2 +- docs/source/scrapers/benchmarks.rst | 23 ------ 3 files changed, 71 insertions(+), 33 deletions(-) delete mode 100644 docs/source/scrapers/benchmarks.rst diff --git a/docs/source/introduction/overview.rst b/docs/source/introduction/overview.rst index 4e2bd604..4dcaadbe 100644 --- a/docs/source/introduction/overview.rst +++ b/docs/source/introduction/overview.rst @@ -30,37 +30,93 @@ ScrapGraphAI supports a wide range of AI models from various providers. Each mod OpenAI Models ------------- - GPT-3.5 Turbo (16,385 tokens) -- GPT-4 (8,192 tokens) +- GPT-3.5 (4,096 tokens) +- GPT-3.5 Turbo Instruct (4,096 tokens) - GPT-4 Turbo Preview (128,000 tokens) -- GPT-4o (128000 tokens) -- GTP-4o-mini (128000 tokens) +- GPT-4 Vision Preview (128,000 tokens) +- GPT-4 (8,192 tokens) +- GPT-4 32k (32,768 tokens) +- GPT-4o (128,000 tokens) +- O1 Preview (128,000 tokens) +- O1 Mini (128,000 tokens) Azure OpenAI Models ------------------- - GPT-3.5 Turbo (16,385 tokens) -- GPT-4 (8,192 tokens) +- GPT-3.5 (4,096 tokens) - GPT-4 Turbo Preview (128,000 tokens) -- GPT-4o (128000 tokens) -- GTP-4o-mini (128000 tokens) +- GPT-4 (8,192 tokens) +- GPT-4 32k (32,768 tokens) +- GPT-4o (128,000 tokens) +- O1 Preview (128,000 tokens) +- O1 Mini (128,000 tokens) Google AI Models ---------------- - Gemini Pro (128,000 tokens) +- Gemini 1.5 Flash (128,000 tokens) - Gemini 1.5 Pro (128,000 tokens) +- Gemini 1.0 Pro (128,000 tokens) Anthropic Models ---------------- - Claude Instant (100,000 tokens) -- Claude 2 (200,000 tokens) +- Claude 2 (9,000 tokens) +- Claude 2.1 (200,000 tokens) - Claude 3 (200,000 tokens) +- Claude 3.5 (200,000 tokens) +- Claude 3 Opus (200,000 tokens) +- Claude 3 Sonnet (200,000 tokens) +- Claude 3 Haiku (200,000 tokens) Mistral AI Models ----------------- -- Mistral Large (128,000 tokens) +- Mistral Large Latest (128,000 tokens) +- Open Mistral Nemo (128,000 tokens) +- Codestral Latest (32,000 tokens) - Open Mistral 7B (32,000 tokens) - Open Mixtral 8x7B (32,000 tokens) +- Open Mixtral 8x22B (64,000 tokens) +- Open Codestral Mamba (256,000 tokens) -For a complete list of supported models and their token limits, please refer to the API documentation. +Ollama Models +------------- +- Command-R (12,800 tokens) +- CodeLlama (16,000 tokens) +- DBRX (32,768 tokens) +- DeepSeek Coder 33B (16,000 tokens) +- Llama2 Series (4,096 tokens) +- Llama3 Series (8,192-128,000 tokens) +- Mistral Models (32,000-128,000 tokens) +- Mixtral 8x22B Instruct (65,536 tokens) +- Phi3 Series (12,800-128,000 tokens) +- Qwen Series (32,000 tokens) + +Hugging Face Models +------------------ +- Grok-1 (8,192 tokens) +- Meta Llama 3 Series (8,192 tokens) +- Google Gemma Series (8,192 tokens) +- Microsoft Phi Series (2,048-131,072 tokens) +- GPT-2 Series (1,024 tokens) +- DeepSeek V2 Series (131,072 tokens) + +Bedrock Models +------------- +- Claude 3 Series (200,000 tokens) +- Llama2 & Llama3 Series (4,096-8,192 tokens) +- Mistral Series (32,768 tokens) +- Titan Embed Text (8,000 tokens) +- Cohere Embed (512 tokens) + +Fireworks Models +--------------- +- Llama V2 7B (4,096 tokens) +- Mixtral 8x7B Instruct (4,096 tokens) +- Llama 3.1 Series (131,072 tokens) +- Mixtral MoE Series (65,536 tokens) + +For a complete and up-to-date list of supported models and their token limits, please refer to the API documentation. Understanding token limits is crucial for optimizing your scraping tasks. Larger token limits allow for processing more text in a single API call, which can be beneficial for scraping lengthy web pages or documents. @@ -139,3 +195,8 @@ Sponsors :width: 15% :alt: Stat Proxies :target: https://dashboard.statproxies.com/?refferal=scrapegraph + +.. image:: ../../assets/scrapedo.png + :width: 11% + :alt: Scrapedo + :target: https://scrape.do diff --git a/docs/source/modules/scrapegraphai.helpers.models_tokens.rst b/docs/source/modules/scrapegraphai.helpers.models_tokens.rst index 173e1bc3..82615b3b 100644 --- a/docs/source/modules/scrapegraphai.helpers.models_tokens.rst +++ b/docs/source/modules/scrapegraphai.helpers.models_tokens.rst @@ -19,7 +19,7 @@ Example usage: print(f"GPT-4 token limit: {gpt4_limit}") # Check the token limit for a specific model - model_name = "gpt-3.5-turbo" + model_name = "gpt-4o-mini" if model_name in models_tokens['openai']: print(f"{model_name} token limit: {models_tokens['openai'][model_name]}") else: diff --git a/docs/source/scrapers/benchmarks.rst b/docs/source/scrapers/benchmarks.rst deleted file mode 100644 index b5521ef1..00000000 --- a/docs/source/scrapers/benchmarks.rst +++ /dev/null @@ -1,23 +0,0 @@ -Benchmarks -========== - -SearchGraph -^^^^^^^^^^^ - -`SearchGraph` instantiates multiple `SmartScraperGraph` object for each URL and extract the data from the HTML. -A concurrent approach is used to speed up the process and the following table shows the time required for a scraping task with different **batch sizes**. -Only two results are taken into account. - -.. list-table:: SearchGraph - :header-rows: 1 - - * - Batch Size - - Total Time (s) - * - 1 - - 31.1 - * - 2 - - 33.52 - * - 4 - - 28.47 - * - 16 - - 21.80