Government Of India
A-
A
A+
microsoft/Phi-3-vision-128k-instruct
Phi-3-vision-128k-instruct is a state-of-the-art multimodal model by Microsoft, designed to process both text and visual inputs with a context length of up to 128,000 tokens.
About Model
As part of the Phi-3 model family, Phi-3-vision-128k-instruct combines text and vision modalities to perform complex reasoning tasks across large contexts. The model is trained on high-quality, reasoning-rich datasets, including synthetic data and filtered publicly available web content. It has undergone supervised fine-tuning and direct preference optimization to enhance instruction-following accuracy and safety. Phi-3-vision-128k-instruct excels in tasks such as image captioning, visual question answering, and document analysis. The model is available on platforms like Hugging Face and Azure AI Foundry, supporting applications that require deep multimodal comprehension.
microsoft/Phi-3-vision-128k-instruct
Metadata
MIT
Microsoft
Multimodal Language Model
N.A.
Open
Sector Agnostic
12/03/25 06:35:30
0
Activity Overview
0
9 
0 
211
Tags
- Reasoning
- Transformers
- NLP
- Microsoft
- Multimodal
- Text Generation
- Visual Question Answering
- Image-to-Text
- Long Context
License Control
MIT
More Models from Microsoft Corporation (India) Pvt. Ltd.
Phi-1.5 - Lightweight Transformer Model for Text Generation and Coding
A 1.3B parameter Transformer model trained on structured QA, NLP tasks, and Python code, optimized for text generation, summarization, and creative writing, with no fine-tuning from human feedback.
Transformers
Code Generation
Instruction Following
Reasoning
NLP
Microsoft
- 0
- 23
- 0
1,394
Updated 1 day(s) ago
Phi-3-Mini-4K-Instruct - Lightweight AI Model for Reasoning and Efficient AI Applications
A 3.8B parameter AI model optimized for memory-efficient reasoning, instruction following, and long-context processing, designed for compute-constrained environments and latency-sensitive applications.
Text Generation
NLP
Microsoft
Transformers
Instruction Following
Reasoning
- 0
- 6
- 0
- 142
Updated 1 day(s) ago
Phi-4-Mini-Instruct - Lightweight Multilingual AI Model
A lightweight AI model optimized for multilingual text generation, reasoning, and instruction adherence, designed for memory-efficient applications and generative AI tasks.
Math
Instruction Following
Function Calling
Coding
Transformers
Microsoft
NLP
Text Generation
Reasoning
- 0
- 10
- 0
- 288
Updated 1 day(s) ago
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.
TAPEX
Transformers
DataRetrieval
SQLExecution
NeuralExecutor
PreTrainedModel
BART
- 0
- 6
- 0
- 249
Updated 11 month(s) ago
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.
TableQuestionAnswering
PreTrainedModel
FactVerification
LargeModel
BART
- 0
- 5
- 0
- 194
Updated 11 month(s) ago
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.
Transformers
BART
NaturalLanguageProcessing
FineTunedModel
TAPEX
FactVerification
DataValidation
TabFact
- 0
- 6
- 0
- 161
Updated 11 month(s) ago
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.
Transformers
BART
NaturalLanguageProcessing
FineTunedModel
TAPEX
DataExtraction
TableQuestionAnswering
WikiTableQuestions
- 0
- 2
- 0
- 230
Updated 11 month(s) ago
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.
TableQuestionAnswering
TabularData
FactVerification
PreTrainedModel
BART
- 0
- 3
- 0
- 167
Updated 11 month(s) ago
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.
WikiTableQuestions
Transformers
BART
NaturalLanguageProcessing
FineTunedModel
TAPEX
DataExtraction
TableQuestionAnswering
- 0
- 3
- 0
- 181
Updated 11 month(s) ago
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.
DataRetrieval
Transformers
BART
NaturalLanguageProcessing
FineTunedModel
TAPEX
WikiSQL
SQLQueryGeneration
- 0
- 5
- 0
- 179
Updated 11 month(s) ago
© 2026 - Copyright AIKosh. All rights reserved.