Mengzi Model
Mengzi pre-training model is a large-scale pre-training language model developed by Langboat team. It can process multi-language and multi-modal data, support a variety of understanding and generation tasks, and can quickly meet the needs of different fields and different application scenarios.
Product Advantages
Support Multiple Model Architectures
- Autoregressive models: such as GPT
- Self-encoding models: such as BERT
- Encoder-Decoder model: T5
Lightweight Model Performance Enhancement
- Fusion of multiple pre-training tasks
- SMART adverserial training
- Knowledge distillation
Knowledge Graph Based Enhancement
- Enhancements with entity extraction
- Knowledge graph enhancement (isa relationship)
- Knowledge graph to text conversion
Linguistic Knowledge Based Enhancement
- Mask mechanism enhanced by syntactic information
- Semantic role embedding enhancement
- Attention weight constrained pruning of dependencies
Few-Shot/Zero-Shot Learning
- Prompt template construction
- Multi-task learning technique
- Common information extraction scenarios, out of the box
Retrieval Based Enhancement
- Knowledge decoupling
- Strong interpretability
- External knowledge components are updated in real time
Product List
Mengzi Generative Pre-trained Transformer Model (Mengzi GPT)
Mengzi Generative Pre-trained Transformer Model (Mengzi-GPT) is a controllable large-scale language model designed for generation scenarios. It can assist users in completing various tasks in specific scenarios through multi-turn interactions.
Learn MoreMengzi-BERT-base
It is suitable for natural language understanding tasks such as text classification, entity recognition, relationship extraction, and reading comprehension.
Learn MoreMengzi-BERT-base-fin
Training based on financial corpus can better complete tasks in financial scenarios.
Learn MoreMengzi-T5-base
It is suitable for controllable text generation tasks such as copywriting generation and news generation.
Learn MoreMengzi-T5-base-MT
Based on the Mengzi-T5-base multi-task model, various tasks can be completed through prompts without training.
Learn MoreMengzi-GPT-neo-base
Based on Chinese corpus training from scratch, it is suitable for tasks such as text continuation and novel generation.
Learn MoreGuohua Diffusion
The Chinese painting-style text-to-image generation model can complete different generation tasks such as movie posters, album covers, and landscape paintings.
Learn MoreSentence Embedding
Based on Langboat's self-developed Mengzi lightweight technology system, it outputs a vector representation of sentence-level text, which can capture the semantic similarity between texts. It is suitable for clustering, regression, anomaly detection, visualization and other tasks.
Learn MoreLarge-scale Model Customization
Based on the Mengzi pre-training model technology developed by Langboat, it provides enterprise customers with faster, more effective and low-cost pre-training model customized training, optimization and deployment services.
Learn MoreBusiness Cooperation Email
Address
11F, Block A, Dinghao DH3 Building, No.3 Haidian Street, Haidian District, Beijing, China
Business Cooperation:
bd@langboat.com
Address:
11F, Block A, Dinghao DH3 Building, No.3 Haidian Street, Haidian District, Beijing, China
Official Accounts:
