Skip to content

Commit

Permalink
Update ntbk
Browse files Browse the repository at this point in the history
  • Loading branch information
@rlancemartin
rlancemartin committed Sep 18, 2024
1 parent a35f56f commit 9e8bc1b
Showing 1 changed file with 114 additions and 44 deletions.
158 changes: 114 additions & 44 deletions ntbk/testing.ipynb
View file
Original file line number Diff line number Diff line change
Expand Up @@ -37,8 +37,7 @@
"outputs": [],
"source": [
"from langgraph_sdk import get_client\n",
"\n",
"client = get_client(url=\"http://localhost:53394\")"
"client = get_client(url=\"http://localhost:56396\")"
]
},
{
Expand All @@ -56,27 +55,41 @@
"metadata": {},
"outputs": [],
"source": [
"schema = {\n",
"schema = {\n",
" \"type\": \"object\",\n",
" \"properties\": {\n",
" \"companies\": {\n",
" \"type\": \"string\",\n",
" \"description\": \"Names of top chip providers for LLM training\",\n",
" },\n",
" \"technologies\": {\n",
" \"type\": \"string\",\n",
" \"description\": \"Brief summary of key chip technologies used for LLM training\",\n",
" },\n",
" \"market_share\": {\n",
" \"type\": \"string\",\n",
" \"description\": \"Overview of market share distribution among top providers\",\n",
" },\n",
" \"future_outlook\": {\n",
" \"type\": \"string\",\n",
" \"description\": \"Brief summary of future prospects and developments in the field\",\n",
" },\n",
" \"type\": \"array\",\n",
" \"items\": {\n",
" \"type\": \"object\",\n",
" \"properties\": {\n",
" \"name\": {\n",
" \"type\": \"string\",\n",
" \"description\": \"Company name\"\n",
" },\n",
" \"technologies\": {\n",
" \"type\": \"string\",\n",
" \"description\": \"Brief summary of key technologies used by the company\"\n",
" },\n",
" \"market_share\": {\n",
" \"type\": \"string\",\n",
" \"description\": \"Overview of market share for this company\"\n",
" },\n",
" \"future_outlook\": {\n",
" \"type\": \"string\",\n",
" \"description\": \"Brief summary of future prospects and developments in the field for this company\"\n",
" },\n",
" \"key_powers\": {\n",
" \"type\": \"string\",\n",
" \"description\": \"Which of the 7 Powers (Scale Economies, Network Economies, Counter Positioning, Switching Costs, Branding, Cornered Resource, Process Power) best describe this company's competitive advantage\"\n",
" }\n",
" },\n",
" \"required\": [\"name\", \"technologies\", \"market_share\", \"future_outlook\"]\n",
" },\n",
" \"description\": \"List of companies\"\n",
" }\n",
" },\n",
" \"required\": [\"companies\", \"technologies\", \"market_share\", \"future_outlook\"],\n",
" \"required\": [\"companies\"]\n",
"}"
]
},
Expand Down Expand Up @@ -105,7 +118,7 @@
"# Stream\n",
"async for event in client.runs.stream(\n",
" thread[\"thread_id\"],\n",
" assistant_id=\"agent\",\n",
" assistant_id= \"agent\",\n",
" input={\n",
" \"topic\": topic,\n",
" \"extraction_schema\": schema,\n",
Expand All @@ -121,7 +134,7 @@
},
{
"cell_type": "code",
"execution_count": 11,
"execution_count": 5,
"metadata": {},
"outputs": [],
"source": [
Expand All @@ -130,26 +143,74 @@
},
{
"cell_type": "code",
"execution_count": 12,
"execution_count": 8,
"metadata": {},
"outputs": [
{
"data": {
"text/markdown": [
"\n",
"# Top 5 Chip Providers for LLM Training\n",
"\n",
"## Companies\n",
"Nvidia, AMD, Google, OpenAI, Microsoft, Meta, Amazon, Baidu, Alibaba, Huawei\n",
"\n",
"## Key Technologies\n",
"GPUs (particularly Nvidia A100 and H100), TPUs (Google's custom AI chips), Large Language Models (LLMs)\n",
"## NVIDIA\n",
"\n",
"**Key Technologies:** GPU accelerators (A100, H100, GH200, Blackwell B200)\n",
"\n",
"**Market Share:** Over 80% of the high-end AI chip market\n",
"\n",
"**Key Powers:** Scale Economies, Network Economies, Branding\n",
"\n",
"**Future Outlook:** Strong growth potential with continued innovation in AI-specific hardware, including the recently announced Blackwell B200 GPU\n",
"\n",
"---\n",
"\n",
"## AMD\n",
"\n",
"**Key Technologies:** MI250 and MI300X GPUs\n",
"\n",
"**Market Share:** Growing presence in AI chip market, with performance reportedly close to NVIDIA in some benchmarks\n",
"\n",
"**Key Powers:** Scale Economies, Counter Positioning\n",
"\n",
"**Future Outlook:** Accelerating roadmap development to compete in the AI market, potential for significant growth\n",
"\n",
"---\n",
"\n",
"## Intel\n",
"\n",
"**Key Technologies:** Gaudi2 and upcoming Gaudi3 AI accelerators\n",
"\n",
"**Market Share:** Emerging player in AI chip market, with Gaudi3 potentially competitive with NVIDIA's H100\n",
"\n",
"**Key Powers:** Scale Economies, Process Power\n",
"\n",
"**Future Outlook:** Promising outlook with Gaudi3 chips, aiming to offer performance comparable to NVIDIA's offerings at a potentially lower cost\n",
"\n",
"---\n",
"\n",
"## Google\n",
"\n",
"**Key Technologies:** Tensor Processing Units (TPUs) v4 and v5e\n",
"\n",
"**Market Share:** Significant player for in-house use and Google Cloud customers\n",
"\n",
"## Market Share\n",
"Nvidia dominates the market, particularly in data center GPUs. In 2023, the top five LLM developers acquired around 88.22% of the market revenue. The global LLM market is projected to grow from $1,590 million in 2023 to $259,800 million in 2030, with a CAGR of 79.80%.\n",
"**Key Powers:** Network Economies, Process Power, Scale Economies\n",
"\n",
"## Future Outlook\n",
"The LLM market is expected to experience explosive growth, with projections reaching $35.43 billion by 2030 at a CAGR of 35.9%. Key developments include advanced pre-training techniques, multimodal models like Google's VideoPoet, and a focus on efficiency and specialized models. The industry is also seeing increased investment in AI infrastructure, prompt engineering, and MLOps (Machine Learning Operations) to support the growing demand for AI technologies.\n"
"**Future Outlook:** Continued development of TPUs with v5e offering up to 4x greater performance per dollar for inference compared to alternatives\n",
"\n",
"---\n",
"\n",
"## Groq\n",
"\n",
"**Key Technologies:** Language Processing Unit (LPU) Inference Engine\n",
"\n",
"**Market Share:** Emerging player focusing on inference, gaining attention for high-speed performance\n",
"\n",
"**Key Powers:** Counter Positioning, Process Power\n",
"\n",
"**Future Outlook:** Potential for rapid adoption among startups due to speed and cost-effectiveness in LLM inference, claiming to serve Mixtral at nearly 500 tokens/second\n",
"\n",
"---\n"
],
"text/plain": [
"<IPython.core.display.Markdown object>"
Expand All @@ -162,26 +223,26 @@
"source": [
"from IPython.display import Markdown, display\n",
"\n",
"\n",
"def format_llm_chip_info(data):\n",
" markdown_text = f\"\"\"\n",
"# Top 5 Chip Providers for LLM Training\n",
" markdown_text = \"# Top 5 Chip Providers for LLM Training\\n\\n\"\n",
" \n",
" for company in data['companies']:\n",
" markdown_text += f\"\"\"\n",
"## {company['name']}\n",
"\n",
"**Key Technologies:** {company['technologies']}\n",
"\n",
"## Companies\n",
"{', '.join(data['companies'].split(', '))}\n",
"**Market Share:** {company['market_share']}\n",
"\n",
"## Key Technologies\n",
"{data['technologies']}\n",
"**Key Powers:** {company.get('key_powers', 'Not specified')}\n",
"\n",
"## Market Share\n",
"{data['market_share']}\n",
"**Future Outlook:** {company['future_outlook']}\n",
"\n",
"## Future Outlook\n",
"{data['future_outlook']}\n",
"---\n",
"\"\"\"\n",
" \n",
" return Markdown(markdown_text)\n",
"\n",
"\n",
"# Display the formatted markdown\n",
"display(format_llm_chip_info(current_state[\"values\"][\"info\"]))"
]
Expand All @@ -198,8 +259,17 @@
]
},
{
"cell_type": "markdown",
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": []
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": []
}
],
Expand Down

0 comments on commit 9e8bc1b

Please sign in to comment.