Government Of India
A-
A
A+
Phi-3-Medium-4K-Instruct ONNX-DirectML - Optimized AI Model for Windows GPU Inference
An ONNX-optimized version of Phi-3-Medium-4K-Instruct, designed for efficient AI inference on Windows machines using DirectML, supporting INT4 quantization for high-performance execution on AMD, Intel, and NVIDIA GPUs.
About Model
Phi-3-Medium-4K-Instruct ONNX-DirectML is a high-performance AI model from Microsoft, optimized for fast execution on Windows-based GPUs using DirectML. This version is quantized to INT4 precision for efficient and scalable inference, allowing deployment across AMD, Intel, and NVIDIA GPUs while maintaining the 4K and 128K token context lengths for structured reasoning and instruction-following tasks.
Phi-3-Medium-4K-Instruct ONNX-DirectML - Optimized AI Model for Windows GPU Inference
Metadata
MIT
Microsoft
Text Generation
N.A.
Open
Sector Agnostic
12/03/25 06:35:40
0
Activity Overview
0
0 
0 
111
Tags
- Text Generation
- Instruction Following
- Reasoning
- ONNX
- DirectML
- Transformers
- NLP
- Microsoft
License Control
MIT
More Models from Microsoft Corporation (India) Pvt. Ltd.
TAPEX: Large SQL Execution Model (Table Pre-training via Learning a Neural SQL Executor)
A large-sized TAPEX model pre-trained to simulate neural SQL execution, enabling the execution of SQL queries on given tables.
Transformers
SQLExecution
PreTrainedModel
TAPEX
DataRetrieval
NeuralExecutor
BART
- 0
- 6
- 0
194
Updated 9 month(s) ago
MICROSOFT CORPORATION (INDIA) PVT. LTD.
TAPEX: Large Model (Table Pre-training via Learning a Neural SQL Executor)
A large-sized pre-trained model designed to enhance table-based question answering and fact verification tasks.
BART
TableQuestionAnswering
FactVerification
PreTrainedModel
LargeModel
- 0
- 5
- 0
- 151
Updated 9 month(s) ago
MICROSOFT CORPORATION (INDIA) PVT. LTD.
TAPEX: TabFact Data enabled Large Finetuned (Table Pre-training via Learning a Neural SQL Executor) Model
A large-sized TAPEX model fine-tuned on the TabFact dataset, designed to enhance performance in table-based fact verification tasks.
FactVerification
NaturalLanguageProcessing
Transformers
BART
DataValidation
FineTunedModel
TabFact
TAPEX
- 0
- 5
- 0
- 114
Updated 9 month(s) ago
MICROSOFT CORPORATION (INDIA) PVT. LTD.
TAPEX (Table Pre-training via Learning a Neural SQL Executor) Large Finetuned Model
A large-sized TAPEX model fine-tuned on the WikiTableQuestions dataset, designed to enhance performance in table-based question answering tasks.
TAPEX
TableQuestionAnswering
NaturalLanguageProcessing
Transformers
BART
DataExtraction
FineTunedModel
WikiTableQuestions
- 0
- 2
- 0
- 164
Updated 9 month(s) ago
MICROSOFT CORPORATION (INDIA) PVT. LTD.
TAPEX: Base Model (Table Pre-training via Learning a Neural SQL Executor)
A base-sized pre-trained model designed to enhance table-based question answering and fact verification tasks.
BART
TableQuestionAnswering
FactVerification
PreTrainedModel
TabularData
- 0
- 3
- 0
- 111
Updated 9 month(s) ago
MICROSOFT CORPORATION (INDIA) PVT. LTD.
TAPEX: WikiTable Questions Data enabled Base Finetuned (Table Pre-training via Learning a Neural SQL Executor) Model
A base-sized TAPEX model fine-tuned on the WikiTableQuestions dataset, designed to enhance performance in table-based question answering tasks.
NaturalLanguageProcessing
TableQuestionAnswering
TAPEX
WikiTableQuestions
FineTunedModel
DataExtraction
BART
Transformers
- 0
- 3
- 0
- 141
Updated 9 month(s) ago
MICROSOFT CORPORATION (INDIA) PVT. LTD.
TAPEX: WikiSQL Data enabled Base Finetuned (Table Pre-training via Learning a Neural SQL Executor) Model
A large-sized TAPEX model fine-tuned on the WikiSQL dataset, optimized for translating natural language questions into SQL queries for effective table-based question answering.
Transformers
NaturalLanguageProcessing
SQLQueryGeneration
TAPEX
WikiSQL
FineTunedModel
DataRetrieval
BART
- 0
- 5
- 0
- 122
Updated 9 month(s) ago
MICROSOFT CORPORATION (INDIA) PVT. LTD.
TAPEX: TabFact Data enabled Base Finetuned (Table Pre-training via Learning a Neural SQL Executor) Model
A base-sized TAPEX model fine-tuned on the TabFact dataset, tailored for verifying the factual accuracy of textual statements against tabular data.
FactVerification
TAPEX
TabFact
FineTunedModel
DataValidation
BART
Transformers
NaturalLanguageProcessing
- 0
- 4
- 0
- 110
Updated 9 month(s) ago
MICROSOFT CORPORATION (INDIA) PVT. LTD.
TAPEX: Base Finetuned (Table Pre-training via Learning a Neural SQL Executor) Model
A base-sized TAPEX model fine-tuned on the WikiSQL dataset, designed to enhance performance in table-based question answering tasks.
DataExtraction
NaturalLanguageProcessing
Transformers
BART
TableQuestionAnswering
FineTunedModel
WikiSQL
TAPEX
- 0
- 5
- 0
- 154
Updated 9 month(s) ago
MICROSOFT CORPORATION (INDIA) PVT. LTD.
BiomedBERT - Domain-Specific Biomedical Language Model
A biomedical NLP model pre-trained from scratch on abstracts and full-text articles from PubMed and PubMed Central, achieving state-of-the-art performance on biomedical language understanding tasks.
Transformers
inference endpoints
exbert
Bert
English
JAX
PyTorch
Fill-Mask
- 0
- 90
- 0
- 1,383
Updated 1 year(s) ago
MICROSOFT CORPORATION (INDIA) PVT. LTD.
© 2026 - Copyright AIKosh. All rights reserved. This portal is developed by National e-Governance Division for AIKosh mission.