GGUF
conversational
A-LLM commited on
Commit
8db9dd5
·
verified ·
1 Parent(s): 8f28e3b

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +22 -3
README.md CHANGED
@@ -1,3 +1,22 @@
1
- ---
2
- license: apache-2.0
3
- ---
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: apache-2.0
3
+ ---
4
+
5
+
6
+ Model Card for Allm-7B-Instruct-v0.1
7
+ Model Overview
8
+ Model Name: Allm-7B-Instruct-v0.1
9
+ Model Type: Large Language Model (LLM)
10
+ Task: Instruction-based Language Generation
11
+ Version: v0.1
12
+ Date: [Release Date]
13
+
14
+ Description:
15
+ Allm-7B-Instruct-v0.1 is a large language model designed to perform tasks based on instructions. This version introduces several enhancements over its predecessor, including a 32k context window (upgraded from the 8k context in the previous version), a Rope-theta value of 1e6, and the elimination of Sliding-Window Attention. For a detailed explanation of these features and the model’s capabilities, please refer to our [paper] and [release blog post].
16
+
17
+ Instruction Format
18
+ To utilize the instruction fine-tuning feature of this model, wrap your prompts with [INST] and [/INST] tokens. Ensure that the very first instruction starts with a sentence ID, while subsequent instructions should not. The response from the assistant will be concluded with an end-of-sentence token ID.
19
+
20
+ Limitations
21
+ The Allm-7B-Instruct model serves as an initial demonstration of how the base model can be fine-tuned to deliver effective instruction-based performance. However, it currently lacks moderation mechanisms. We are eager to collaborate with the community to develop and implement guardrails that ensure the model produces moderated outputs, making it suitable for deployment in regulated environments.
22
+