Stablelm 3B 4e1t is an open-source language model by stabilityai. Features: 3b LLM, VRAM: 5.6GB, Context: 4K, License: cc-by-sa-4.0, LLM Explorer Score: 0.21, Arc: 46.6, HellaSwag: 75.9, MMLU: 45.2, GSM8K: 3.3.
Stablelm 3B 4e1t Parameters and Internals
Model Type auto-regressive, transformer, decoder-only
Use Cases
Primary Use Cases: Limitations: May exhibit unreliable, unsafe, or undesirable behaviors requiring correction through evaluation and fine-tuning, Dataset may contain offensive or inappropriate content, Exercise caution for production systems
Considerations: Not suitable for applications that may cause deliberate or unintentional harm.
Additional Notes Recommended to fine-tune the base StableLM-3B-4E1T for downstream tasks.
Supported Languages Training Details
Data Sources: tiiuae/falcon-refinedweb, togethercomputer/RedPajama-Data-1T, CarperAI/pilev2-dev, bigcode/starcoderdata, allenai/peS2o
Data Volume: Methodology: Pre-trained in bfloat16 precision, optimized with AdamW, trained using the NeoX tokenizer with a vocabulary size of 50,257
Context Length: Training Time: Hardware Used: 256 NVIDIA A100 40GB GPUs (AWS P4d instances)
Model Architecture: decoder-only transformer, similar to LLaMA, with modifications in position embeddings, normalization, and tokenizer using GPT-NeoX
Responsible Ai Considerations
Mitigation Strategies: Users must evaluate and fine-tune the model for safe performance.
Rank the Stablelm 3B 4e1t capabilities
Have you tried this model? Rate its performance — this feedback helps the ML community find the right model for their needs.
Instruction Following and Task Automation
Factuality and Completeness of Knowledge
Censorship and Alignment
Data Analysis and Insight Generation
Text Generation
Text Summarization and Feature Extraction
Code Generation
Multi-Language Support and Translation
Quantized Models of the Stablelm 3B 4e1t
Best Alternatives to Stablelm 3B 4e1t
Note: green Score (e.g. "73.2 ") means that the model is better than stabilityai/stablelm-3b-4e1t .
Expand