Upload 3 files

Add hardware data

hardware_2026.json

@@ -0,0 +1,177 @@
+{
+  "version": "2026.1.0",
+  "last_updated": "2026-03-15",
+  "citation": "https://www.iaexplore.com/html/en/blog/ai/npu-tpu-gpu-ai-hardware-guide.html",
+  "hardware": {
+    "rtx-5090-32gb": {
+      "id": "rtx-5090-32gb",
+      "name": "NVIDIA GeForce RTX 5090",
+      "vendor": "NVIDIA",
+      "category": "discrete_gpu",
+      "vram_gb": 32.0,
+      "usable_vram_gb": 31.2,
+      "bandwidth_gb_s": 1792.0,
+      "tflops_fp16": 209.0,
+      "tops_int8": 418.0,
+      "tdp_watts": 575,
+      "bus_type": "GDDR7 512-bit",
+      "efficiency_factor": 0.72,
+      "reference_url": "https://www.iaexplore.com/html/en/blog/ai/npu-tpu-gpu-ai-hardware-guide.html",
+      "benchmark_guide_es": "https://www.iaexplore.com/html/es/blog/comparativas/los-mejores-ordenadores-que-pc-comprar-segun-tu-uso-y-presupuesto.html",
+      "benchmark_guide_en": "https://www.iaexplore.com/html/en/blog/comparisons/best-computers-which-pc-to-buy-based-on-your-needs-and-budget.html"
+    },
+    "rtx-4090-24gb": {
+      "id": "rtx-4090-24gb",
+      "name": "NVIDIA GeForce RTX 4090",
+      "vendor": "NVIDIA",
+      "category": "discrete_gpu",
+      "vram_gb": 24.0,
+      "usable_vram_gb": 23.4,
+      "bandwidth_gb_s": 1008.0,
+      "tflops_fp16": 165.0,
+      "tops_int8": 330.0,
+      "tdp_watts": 450,
+      "bus_type": "GDDR6X 384-bit",
+      "efficiency_factor": 0.72,
+      "reference_url": "https://www.iaexplore.com/html/en/blog/ai/npu-tpu-gpu-ai-hardware-guide.html",
+      "benchmark_guide_es": "https://www.iaexplore.com/html/es/blog/comparativas/los-mejores-ordenadores-que-pc-comprar-segun-tu-uso-y-presupuesto.html",
+      "benchmark_guide_en": "https://www.iaexplore.com/html/en/blog/comparisons/best-computers-which-pc-to-buy-based-on-your-needs-and-budget.html"
+    },
+    "rtx-4070-super-12gb": {
+      "id": "rtx-4070-super-12gb",
+      "name": "NVIDIA GeForce RTX 4070 SUPER",
+      "vendor": "NVIDIA",
+      "category": "discrete_gpu",
+      "vram_gb": 12.0,
+      "usable_vram_gb": 11.5,
+      "bandwidth_gb_s": 504.0,
+      "tflops_fp16": 71.0,
+      "tops_int8": 142.0,
+      "tdp_watts": 220,
+      "bus_type": "GDDR6X 192-bit",
+      "efficiency_factor": 0.72,
+      "reference_url": "https://www.iaexplore.com/html/en/blog/ai/npu-tpu-gpu-ai-hardware-guide.html",
+      "benchmark_guide_es": "https://www.iaexplore.com/html/es/blog/comparativas/los-mejores-ordenadores-que-pc-comprar-segun-tu-uso-y-presupuesto.html",
+      "benchmark_guide_en": "https://www.iaexplore.com/html/en/blog/comparisons/best-computers-which-pc-to-buy-based-on-your-needs-and-budget.html"
+    },
+    "rtx-4060-laptop-8gb": {
+      "id": "rtx-4060-laptop-8gb",
+      "name": "NVIDIA GeForce RTX 4060 Mobile",
+      "vendor": "NVIDIA",
+      "category": "mobile_gpu",
+      "vram_gb": 8.0,
+      "usable_vram_gb": 7.6,
+      "bandwidth_gb_s": 256.0,
+      "tflops_fp16": 30.0,
+      "tops_int8": 60.0,
+      "tdp_watts": 115,
+      "bus_type": "GDDR6 128-bit",
+      "efficiency_factor": 0.72,
+      "reference_url": "https://www.iaexplore.com/html/en/blog/ai/npu-tpu-gpu-ai-hardware-guide.html",
+      "benchmark_guide_es": "https://www.iaexplore.com/html/es/blog/comparativas/los-mejores-ordenadores-que-pc-comprar-segun-tu-uso-y-presupuesto.html",
+      "benchmark_guide_en": "https://www.iaexplore.com/html/en/blog/comparisons/best-computers-which-pc-to-buy-based-on-your-needs-and-budget.html"
+    },
+    "apple-m4-max-128gb": {
+      "id": "apple-m4-max-128gb",
+      "name": "Apple M4 Max (128GB Unified)",
+      "vendor": "Apple",
+      "category": "unified_memory",
+      "vram_gb": 128.0,
+      "usable_vram_gb": 120.0,
+      "bandwidth_gb_s": 546.0,
+      "tflops_fp16": 54.0,
+      "tops_int8": 108.0,
+      "tdp_watts": 90,
+      "bus_type": "512-bit LPDDR5X-8533",
+      "efficiency_factor": 0.72,
+      "reference_url": "https://www.iaexplore.com/html/en/blog/ai/npu-tpu-gpu-ai-hardware-guide.html",
+      "benchmark_guide_es": "https://www.iaexplore.com/html/es/blog/comparativas/los-mejores-ordenadores-que-pc-comprar-segun-tu-uso-y-presupuesto.html",
+      "benchmark_guide_en": "https://www.iaexplore.com/html/en/blog/comparisons/best-computers-which-pc-to-buy-based-on-your-needs-and-budget.html"
+    },
+    "apple-m4-pro-48gb": {
+      "id": "apple-m4-pro-48gb",
+      "name": "Apple M4 Pro (48GB Unified)",
+      "vendor": "Apple",
+      "category": "unified_memory",
+      "vram_gb": 48.0,
+      "usable_vram_gb": 44.0,
+      "bandwidth_gb_s": 273.0,
+      "tflops_fp16": 32.0,
+      "tops_int8": 64.0,
+      "tdp_watts": 60,
+      "bus_type": "256-bit LPDDR5X-8533",
+      "efficiency_factor": 0.72,
+      "reference_url": "https://www.iaexplore.com/html/en/blog/ai/npu-tpu-gpu-ai-hardware-guide.html",
+      "benchmark_guide_es": "https://www.iaexplore.com/html/es/blog/comparativas/los-mejores-ordenadores-que-pc-comprar-segun-tu-uso-y-presupuesto.html",
+      "benchmark_guide_en": "https://www.iaexplore.com/html/en/blog/comparisons/best-computers-which-pc-to-buy-based-on-your-needs-and-budget.html"
+    },
+    "apple-m4-base-16gb": {
+      "id": "apple-m4-base-16gb",
+      "name": "Apple M4 (16GB Unified)",
+      "vendor": "Apple",
+      "category": "unified_memory",
+      "vram_gb": 16.0,
+      "usable_vram_gb": 13.5,
+      "bandwidth_gb_s": 120.0,
+      "tflops_fp16": 18.0,
+      "tops_int8": 38.0,
+      "tdp_watts": 30,
+      "bus_type": "128-bit LPDDR5X",
+      "efficiency_factor": 0.72,
+      "reference_url": "https://www.iaexplore.com/html/en/blog/ai/npu-tpu-gpu-ai-hardware-guide.html",
+      "benchmark_guide_es": "https://www.iaexplore.com/html/es/blog/comparativas/los-mejores-ordenadores-que-pc-comprar-segun-tu-uso-y-presupuesto.html",
+      "benchmark_guide_en": "https://www.iaexplore.com/html/en/blog/comparisons/best-computers-which-pc-to-buy-based-on-your-needs-and-budget.html"
+    },
+    "amd-ryzen-ai-max-395-128gb": {
+      "id": "amd-ryzen-ai-max-395-128gb",
+      "name": "AMD Ryzen AI Max+ 395 (Strix Halo 128GB)",
+      "vendor": "AMD",
+      "category": "unified_apu_npu",
+      "vram_gb": 96.0,
+      "usable_vram_gb": 92.0,
+      "bandwidth_gb_s": 256.0,
+      "tflops_fp16": 48.0,
+      "tops_int8": 126.0,
+      "tdp_watts": 120,
+      "bus_type": "256-bit LPDDR5X-8000 Unified",
+      "efficiency_factor": 0.60,
+      "reference_url": "https://www.iaexplore.com/html/en/blog/ai/npu-tpu-gpu-ai-hardware-guide.html",
+      "benchmark_guide_es": "https://www.iaexplore.com/html/es/blog/comparativas/los-mejores-ordenadores-que-pc-comprar-segun-tu-uso-y-presupuesto.html",
+      "benchmark_guide_en": "https://www.iaexplore.com/html/en/blog/comparisons/best-computers-which-pc-to-buy-based-on-your-needs-and-budget.html"
+    },
+    "intel-core-ultra-9-288v-32gb": {
+      "id": "intel-core-ultra-9-288v-32gb",
+      "name": "Intel Core Ultra 9 288V (Lunar Lake 32GB)",
+      "vendor": "Intel",
+      "category": "integrated_npu_arc",
+      "vram_gb": 24.0,
+      "usable_vram_gb": 22.0,
+      "bandwidth_gb_s": 136.0,
+      "tflops_fp16": 32.0,
+      "tops_int8": 120.0,
+      "tdp_watts": 30,
+      "bus_type": "On-Package 128-bit LPDDR5X-8533",
+      "efficiency_factor": 0.60,
+      "reference_url": "https://www.iaexplore.com/html/en/blog/ai/npu-tpu-gpu-ai-hardware-guide.html",
+      "benchmark_guide_es": "https://www.iaexplore.com/html/es/blog/comparativas/los-mejores-ordenadores-que-pc-comprar-segun-tu-uso-y-presupuesto.html",
+      "benchmark_guide_en": "https://www.iaexplore.com/html/en/blog/comparisons/best-computers-which-pc-to-buy-based-on-your-needs-and-budget.html"
+    },
+    "qualcomm-snapdragon-x-elite-64gb": {
+      "id": "qualcomm-snapdragon-x-elite-64gb",
+      "name": "Qualcomm Snapdragon X Elite (64GB)",
+      "vendor": "Qualcomm",
+      "category": "unified_soc_npu",
+      "vram_gb": 48.0,
+      "usable_vram_gb": 44.0,
+      "bandwidth_gb_s": 135.0,
+      "tflops_fp16": 15.0,
+      "tops_int8": 45.0,
+      "tdp_watts": 45,
+      "bus_type": "128-bit LPDDR5X-8448",
+      "efficiency_factor": 0.60,
+      "reference_url": "https://www.iaexplore.com/html/pt/blog/comparativos/melhores-tablets-qualidade-preco.html",
+      "benchmark_guide_es": "https://www.iaexplore.com/html/es/blog/comparativas/los-mejores-ordenadores-que-pc-comprar-segun-tu-uso-y-presupuesto.html",
+      "benchmark_guide_en": "https://www.iaexplore.com/html/en/blog/comparisons/best-computers-which-pc-to-buy-based-on-your-needs-and-budget.html"
+    }
+  }
+}

huggingface_README.md

@@ -0,0 +1,141 @@
+---
+annotations_creators:
+- machine-generated
+- expert-generated
+language:
+- en
+- es
+- pt
+license: mit
+multilinguality:
+- multilingual
+size_categories:
+- n<1K
+source_datasets:
+- original
+task_categories:
+- tabular-classification
+- tabular-regression
+- feature-extraction
+task_ids: []
+pretty_name: "2026 LLM Hardware & VRAM Memory Profiling Dataset"
+tags:
+- llm
+- vram
+- inference
+- hardware
+- llama-cpp
+- ollama
+- vllm
+- gqa
+- moe
+- benchmarks
+- iaexplore
+configs:
+- config_name: default
+  data_files:
+  - split: train
+    path:
+    - hardware_2026.json
+    - models_2026.json
+---
+
+# 🧠 IAExplore 2026 LLM Hardware & VRAM Memory Profiling Dataset
+
+[![DOI](https://zenodo.org/badge/DOI/10.5281/zenodo.14889210.svg)](https://zenodo.org/badge/latestdoi/14889210)
+[![Documentation](https://img.shields.io/badge/docs-IAExplore.com-blue.svg)](https://www.iaexplore.com/html/en/blog/ai/npu-tpu-gpu-ai-hardware-guide.html)
+[![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](https://opensource.org/licenses/MIT)
+
+Official dataset providing verified architectural parameters for cutting-edge Large Language Models (LLMs) and 2026 consumer/workstation AI hardware compute profiles.
+
+Curated by the **[IAExplore Research & Systems Infrastructure Team](https://www.iaexplore.com/)**.
+
+---
+
+## 📌 Dataset Summary
+
+This dataset serves as the ground-truth technical foundation for:
+1. **Mathematical VRAM Allocation**: Weight sizing with unquantized metadata overhead factor ($1.08$), KV-Cache memory consumption accounting for Grouped-Query Attention (GQA), and CUDA/Metal runtime execution buffers.
+2. **Mixture-of-Experts (MoE) Bandwidth Profiling**: Distinguishing active parameter footprints ($P_{\text{active}}$) from total stored parameters ($P_{\text{total}}$) per token generation step.
+3. **Hardware Latency & Tokens/Second Estimation**: Exact physical memory bandwidths ($\text{BW}_{\text{mem}}$), bus topologies (GDDR7, LPDDR5X, GDDR6X, HBM), and bus saturation efficiency coefficients ($\eta$).
+
+---
+
+## 🗂️ Data Structure
+
+### 1. `models_2026.json`
+Contains verified tensor specifications for:
+- `llama-3.3-70b` (Meta AI)
+- `llama-3.1-8b` (Meta AI)
+- `deepseek-r1-14b` (DeepSeek AI)
+- `deepseek-r1-32b` (DeepSeek AI)
+- `qwen-2.5-coder-32b` (Alibaba Cloud)
+- `mixtral-8x7b-moe` (Mistral AI)
+- `gemma-2-9b` (Google DeepMind)
+- `phi-4-14b` (Microsoft Research)
+
+Key fields:
+- `total_params_b`: Total billion parameters.
+- `active_params_b`: Active parameters loaded per generation token (vital for MoE architectures).
+- `num_hidden_layers` ($L$): Number of transformer blocks.
+- `num_attention_heads` ($H_{\text{q}}$): Query heads.
+- `num_key_value_heads` ($H_{\text{kv}}$): Key/Value heads for GQA/MHA.
+- `head_dim` ($D_{\text{head}}$): Dimension per head.
+- `gqa_factor`: Compression ratio ($H_{\text{q}} / H_{\text{kv}}$).
+
+### 2. `hardware_2026.json`
+Profiles for 2026 silicon across Discrete GPUs, Apple Silicon, and NPU/APU SoC systems:
+- **NVIDIA GeForce RTX 5090** (32 GB GDDR7, 1792 GB/s, 575W)
+- **NVIDIA GeForce RTX 4090** (24 GB GDDR6X, 1008 GB/s, 450W)
+- **NVIDIA GeForce RTX 4070 SUPER** (12 GB GDDR6X, 504 GB/s, 220W)
+- **NVIDIA GeForce RTX 4060 Mobile** (8 GB GDDR6, 256 GB/s, 115W)
+- **Apple M4 Max** (128 GB Unified LPDDR5X, 546 GB/s, 90W)
+- **Apple M4 Pro** (48 GB Unified LPDDR5X, 273 GB/s, 60W)
+- **Apple M4** (16 GB Unified LPDDR5X, 120 GB/s, 30W)
+- **AMD Ryzen AI Max+ 395 Strix Halo** (96 GB assignable, 256 GB/s, 126 TOPS, 120W)
+- **Intel Core Ultra 9 288V Lunar Lake** (24 GB shared, 136 GB/s, 120 TOPS, 30W)
+- **Qualcomm Snapdragon X Elite** (48 GB assignable, 135 GB/s, 45 TOPS, 45W)
+
+---
+
+## 🔬 Mathematical Formulas
+
+### Model Weights Memory:
+$$\text{VRAM}_{\text{weights}} = \frac{P_{\text{total}} \times 10^9 \times (\text{bits}_{\text{quant}} / 8)}{1024^3} \times 1.08$$
+
+### Exact KV-Cache Memory (with GQA):
+$$\text{VRAM}_{\text{kv}} = \frac{2 \times L \times H_{\text{kv}} \times D_{\text{head}} \times C_{\text{ctx}} \times B_{\text{batch}} \times (\text{bits}_{\text{kv}} / 8)}{1024^3}$$
+
+### Autoregressive Inference Throughput (Tokens/Second):
+$$\text{TPS}_{\text{gen}} = \frac{\text{BW}_{\text{mem}} \times \eta_{\text{efficiency}}}{\text{VRAM}_{\text{active}} + (\text{VRAM}_{\text{kv}} \times 0.15)}$$
+
+---
+
+## 🔗 Authoritative References & In-Depth Guides
+
+- **AI Hardware Architecture Guide (EN)**: [NPU, TPU and GPU Complete Guide](https://www.iaexplore.com/html/en/blog/ai/npu-tpu-gpu-ai-hardware-guide.html)
+- **Guía de Hardware de IA (ES)**: [Arquitecturas NPU, TPU y GPU para Inteligencia Artificial](https://www.iaexplore.com/html/es/blog/ia/npu-tpu-gpu-guia-hardware-inteligencia-artificial.html)
+- **Guia de Hardware de IA (PT)**: [Hardware para IA: Guia Completo NPU, TPU e GPU](https://www.iaexplore.com/html/pt/blog/ia/npu-tpu-gpu-guia-hardware-inteligencia-artificial.html)
+- **2026 PC & Hardware Benchmarks (EN)**: [Best Computers of 2026: Benchmark Guide](https://www.iaexplore.com/html/en/blog/comparisons/best-computers-which-pc-to-buy-based-on-your-needs-and-budget.html)
+- **Comparativas de Ordenadores 2026 (ES)**: [Los Mejores Ordenadores de 2026: Guía y Benchmarks](https://www.iaexplore.com/html/es/blog/comparativas/los-mejores-ordenadores-que-pc-comprar-segun-tu-uso-y-presupuesto.html)
+- **Melhores Computadores 2026 (PT)**: [Melhores Computadores 2026: Guia Completo de Compra](https://www.iaexplore.com/html/pt/blog/comparativos/melhores-computadores-que-pc-comprar-segundo-uso-orcamento.html)
+- **ARM & Mobile AI Devices 2026 (PT)**: [Melhores Tablets e Dispositivos ARM](https://www.iaexplore.com/html/pt/blog/comparativos/melhores-tablets-qualidade-preco.html)
+- **Landauer Principle & Reversible Computing (ES)**: [Física de la Computación y Límite Termodinámico](https://www.iaexplore.com/html/es/blog/tecnologia/computacion-reversible-principio-landauer-calor-procesadores.html)
+- **Interactive Neural Network Simulator**: [IAExplore Real-Time Simulator](https://www.iaexplore.com/neural-network-sim/)
+
+---
+
+## 📜 Citation
+
+If you use this dataset in academic research, benchmark pipelines, or LLM serving infrastructure, please cite:
+
+```bibtex
+@misc{iaexplore_llm_profiler_2026,
+  author = {IAExplore Research Team},
+  title = {IAExplore LLM Profiler: Deterministic VRAM and Hardware Inference Dataset},
+  year = {2026},
+  publisher = {Zenodo},
+  doi = {10.5281/zenodo.14889210},
+  url = {https://www.iaexplore.com/}
+}
+```

models_2026.json

@@ -0,0 +1,161 @@
+{
+  "version": "2026.1.0",
+  "citation": "https://www.iaexplore.com/neural-network-sim/",
+  "documentation": "https://www.iaexplore.com/html/en/blog/ai/npu-tpu-gpu-ai-hardware-guide.html",
+  "models": {
+    "llama-3.3-70b": {
+      "id": "llama-3.3-70b",
+      "name": "Meta Llama 3.3 70B Instruct",
+      "family": "Llama 3.3",
+      "architecture": "dense",
+      "total_params_b": 70.6,
+      "active_params_b": 70.6,
+      "num_hidden_layers": 80,
+      "num_attention_heads": 64,
+      "num_key_value_heads": 8,
+      "head_dim": 128,
+      "hidden_size": 8192,
+      "intermediate_size": 28672,
+      "vocab_size": 128256,
+      "max_context_length": 131072,
+      "default_context": 32768,
+      "gqa_factor": 8.0,
+      "reference_url": "https://www.iaexplore.com/html/en/blog/ai/npu-tpu-gpu-ai-hardware-guide.html"
+    },
+    "llama-3.1-8b": {
+      "id": "llama-3.1-8b",
+      "name": "Meta Llama 3.1 8B Instruct",
+      "family": "Llama 3.1",
+      "architecture": "dense",
+      "total_params_b": 8.03,
+      "active_params_b": 8.03,
+      "num_hidden_layers": 32,
+      "num_attention_heads": 32,
+      "num_key_value_heads": 8,
+      "head_dim": 128,
+      "hidden_size": 4096,
+      "intermediate_size": 14336,
+      "vocab_size": 128256,
+      "max_context_length": 131072,
+      "default_context": 8192,
+      "gqa_factor": 4.0,
+      "reference_url": "https://www.iaexplore.com/html/es/blog/ia/npu-tpu-gpu-guia-hardware-inteligencia-artificial.html"
+    },
+    "deepseek-r1-14b": {
+      "id": "deepseek-r1-14b",
+      "name": "DeepSeek R1 Distill Qwen 14B",
+      "family": "DeepSeek R1",
+      "architecture": "dense",
+      "total_params_b": 14.8,
+      "active_params_b": 14.8,
+      "num_hidden_layers": 48,
+      "num_attention_heads": 40,
+      "num_key_value_heads": 8,
+      "head_dim": 128,
+      "hidden_size": 5120,
+      "intermediate_size": 13824,
+      "vocab_size": 129280,
+      "max_context_length": 131072,
+      "default_context": 32768,
+      "gqa_factor": 5.0,
+      "reference_url": "https://www.iaexplore.com/html/en/blog/ai/npu-tpu-gpu-ai-hardware-guide.html"
+    },
+    "deepseek-r1-32b": {
+      "id": "deepseek-r1-32b",
+      "name": "DeepSeek R1 Distill Qwen 32B",
+      "family": "DeepSeek R1",
+      "architecture": "dense",
+      "total_params_b": 32.8,
+      "active_params_b": 32.8,
+      "num_hidden_layers": 64,
+      "num_attention_heads": 40,
+      "num_key_value_heads": 8,
+      "head_dim": 128,
+      "hidden_size": 5120,
+      "intermediate_size": 27648,
+      "vocab_size": 129280,
+      "max_context_length": 131072,
+      "default_context": 32768,
+      "gqa_factor": 5.0,
+      "reference_url": "https://www.iaexplore.com/html/es/blog/ia/npu-tpu-gpu-guia-hardware-inteligencia-artificial.html"
+    },
+    "qwen-2.5-coder-32b": {
+      "id": "qwen-2.5-coder-32b",
+      "name": "Qwen 2.5 Coder 32B Instruct",
+      "family": "Qwen 2.5",
+      "architecture": "dense",
+      "total_params_b": 32.5,
+      "active_params_b": 32.5,
+      "num_hidden_layers": 64,
+      "num_attention_heads": 40,
+      "num_key_value_heads": 8,
+      "head_dim": 128,
+      "hidden_size": 5120,
+      "intermediate_size": 27648,
+      "vocab_size": 152064,
+      "max_context_length": 131072,
+      "default_context": 32768,
+      "gqa_factor": 5.0,
+      "reference_url": "https://www.iaexplore.com/html/en/blog/ai/npu-tpu-gpu-ai-hardware-guide.html"
+    },
+    "mixtral-8x7b-moe": {
+      "id": "mixtral-8x7b-moe",
+      "name": "Mistral AI Mixtral 8x7B MoE",
+      "family": "Mixtral",
+      "architecture": "moe",
+      "total_params_b": 46.7,
+      "active_params_b": 12.9,
+      "num_hidden_layers": 32,
+      "num_attention_heads": 32,
+      "num_key_value_heads": 8,
+      "head_dim": 128,
+      "hidden_size": 4096,
+      "intermediate_size": 14336,
+      "vocab_size": 32000,
+      "max_context_length": 32768,
+      "default_context": 16384,
+      "gqa_factor": 4.0,
+      "num_experts": 8,
+      "num_experts_per_tok": 2,
+      "reference_url": "https://www.iaexplore.com/html/en/blog/ai/npu-tpu-gpu-ai-hardware-guide.html"
+    },
+    "gemma-2-9b": {
+      "id": "gemma-2-9b",
+      "name": "Google Gemma 2 9B Instruct",
+      "family": "Gemma 2",
+      "architecture": "dense",
+      "total_params_b": 9.24,
+      "active_params_b": 9.24,
+      "num_hidden_layers": 42,
+      "num_attention_heads": 16,
+      "num_key_value_heads": 8,
+      "head_dim": 256,
+      "hidden_size": 3584,
+      "intermediate_size": 14336,
+      "vocab_size": 256000,
+      "max_context_length": 8192,
+      "default_context": 8192,
+      "gqa_factor": 2.0,
+      "reference_url": "https://www.iaexplore.com/html/en/blog/ai/npu-tpu-gpu-ai-hardware-guide.html"
+    },
+    "phi-4-14b": {
+      "id": "phi-4-14b",
+      "name": "Microsoft Phi-4 14B",
+      "family": "Phi-4",
+      "architecture": "dense",
+      "total_params_b": 14.7,
+      "active_params_b": 14.7,
+      "num_hidden_layers": 40,
+      "num_attention_heads": 40,
+      "num_key_value_heads": 10,
+      "head_dim": 128,
+      "hidden_size": 5120,
+      "intermediate_size": 17920,
+      "vocab_size": 100352,
+      "max_context_length": 16384,
+      "default_context": 16384,
+      "gqa_factor": 4.0,
+      "reference_url": "https://www.iaexplore.com/html/es/blog/ia/npu-tpu-gpu-guia-hardware-inteligencia-artificial.html"
+    }
+  }
+}