CraneAILabs/swahili-gemma-1b
1
0
Fork 0
Code Issues Pull Requests Actions Projects Releases Wiki Activity

初始化项目，由ModelHub XC社区提供模型

Browse Source 
Model: CraneAILabs/swahili-gemma-1b
Source: Original Platform
This commit is contained in:
ModelHub XC
2026-05-17 22:37:29 +08:00
commit 354ce0f3a5
15 changed files with 51895 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

17
swahili_results.csv Normal file
View File

@@ -0,0 +1,17 @@
Rank,Model,Type,BLEU,chrF++,Our Model
1,Google Translate,Commercial Service,39.94,66.71,FALSE
2,Gemini 2.0 Flash 001,Google,35.57,64.62,FALSE
3,Claude Sonnet 4,Anthropic,34.93,64.27,FALSE
4,Chatgpt 4o Latest,OpenAI,34.57,64.18,FALSE
5,Gemini 2.5 Flash,Google,34.16,63.51,FALSE
6,Gpt 5 Mini,OpenAI,31.75,62.41,FALSE
7,Llama 4 Scout,Meta,31.46,61.77,FALSE
8,Llama 4 Maverick,Meta,31.35,61.71,FALSE
9,Gpt Oss 120B,OpenAI,30.64,60.47,FALSE
10,Gpt 5 Nano,OpenAI,30.19,61.35,FALSE
11,Gemma 3 27B,Google,29.38,60.02,FALSE
12,Swahili Gemma 1B (Our Model),Specialized Fine-tuned,27.59,56.84,TRUE
13,Gpt Oss 20B,OpenAI,20.13,50.9,FALSE
14,Gemma 3 4B,Google,10.91,44.1,FALSE
15,Gemma 3N E4B,Google,6.27,42.47,FALSE
16,Gemma 3 1B (vLLM Baseline),General Model,2.78,24.79,FALSE
1 Rank Model Type BLEU chrF++ Our Model
2 1 Google Translate Commercial Service 39.94 66.71 FALSE
3 2 Gemini 2.0 Flash 001 Google 35.57 64.62 FALSE
4 3 Claude Sonnet 4 Anthropic 34.93 64.27 FALSE
5 4 Chatgpt 4o Latest OpenAI 34.57 64.18 FALSE
6 5 Gemini 2.5 Flash Google 34.16 63.51 FALSE
7 6 Gpt 5 Mini OpenAI 31.75 62.41 FALSE
8 7 Llama 4 Scout Meta 31.46 61.77 FALSE
9 8 Llama 4 Maverick Meta 31.35 61.71 FALSE
10 9 Gpt Oss 120B OpenAI 30.64 60.47 FALSE
11 10 Gpt 5 Nano OpenAI 30.19 61.35 FALSE
12 11 Gemma 3 27B Google 29.38 60.02 FALSE
13 12 Swahili Gemma 1B (Our Model) Specialized Fine-tuned 27.59 56.84 TRUE
14 13 Gpt Oss 20B OpenAI 20.13 50.9 FALSE
15 14 Gemma 3 4B Google 10.91 44.1 FALSE
16 15 Gemma 3N E4B Google 6.27 42.47 FALSE
17 16 Gemma 3 1B (vLLM Baseline) General Model 2.78 24.79 FALSE