File size: 1,192 Bytes
70adf25
 
 
8be5b01
 
70adf25
 
 
8be5b01
d6733a4
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
---
license: mit
---
This model needs further fine tuning.

See:
txtai-rag.py for rag implimentation with txt ai wikipedia. 

Llamafyd version of Qwen .5B further fine tuned 1 epoch on wiki, math, science, and chat datasets. Based on Cinder data. This model further fine tuned 1 epoch on rag data.

Rough list of final datasets:
formatted_beaugogh-openorca-multiplechoice-10k.txt
formatted_BYC-Sophie-samsum-chatgpt-summary.txt
formatted_conversation_bio.txt
formatted_conversation_create_cinder_1.txt
formatted_conversation_Electrical-engineering.txt
formatted_conversation_multiturn_stem.txt
formatted_conversation_physics.txt
formatted_conversation_qa_rag_chem_prog_dataset.txt
formatted_conversation_qa_robot_ai_dataset.txt
formatted_conversation_qa_shopify_dataset1_rag.txt
formatted_conversation_qa_shopify_dataset_rag.txt
formatted_dyumat-databricks-dolly-5k-rag-split.txt
formatted_Hypoxiic-wikipedia-summary-subset1k-summary_token.txt
formatted_neural-bridge-rag-dataset-12000.txt
formatted_rachid16-rag_finetuning_data.txt
formatted_tiny_stories_1_summary_token_tag_token-xaa.txt
formatted_tiny_stories_2_summary_token_assistant-xah.txt
med_rag_small.txt
z_formatted_cinder_test.txt