- 
	
	
	  lightblue/DeepSeek-R1-Distill-Qwen-1.5B-Multilingual2B • Updated • 121 • 24
- 
	
	
	  lightblue/DeepSeek-R1-Distill-Qwen-7B-Multilingual8B • Updated • 67 • 22
- 
	
	
	  lightblue/DeepSeek-R1-Distill-Qwen-14B-Multilingual15B • Updated • 42 • 13
- 
	
	
	  lightblue/DeepSeek-R1-Distill-Qwen-7B-JapaneseText Generation • 8B • Updated • 19 • • 31
AI & ML interests
None defined yet.
Recent Activity
	View all activity
	
			Organization Card
		
		
 
We are a Tokyo-based firm striving to make the most useful, trustworthy, and reliable AI for our customers and the AI community.
Find out more about us on our website.
- 
	
	
	  lightblue/DeepSeek-R1-Distill-Qwen-1.5B-Multilingual2B • Updated • 121 • 24
- 
	
	
	  lightblue/DeepSeek-R1-Distill-Qwen-7B-Multilingual8B • Updated • 67 • 22
- 
	
	
	  lightblue/DeepSeek-R1-Distill-Qwen-14B-Multilingual15B • Updated • 42 • 13
- 
	
	
	  lightblue/DeepSeek-R1-Distill-Qwen-7B-JapaneseText Generation • 8B • Updated • 19 • • 31
The models trained under our Karasu and Qarasu project
			
	
	
			models
			60
		
			
	
	
	
	
	 
				lightblue/lb-reranker-0.5B-v1.0
			Text Ranking
			• 
		
				0.5B
			• 
	
				Updated
					
				
				• 
					
					269
				
	
				• 
					
					74
				
 
				lightblue/reranker_0.5_bincont_filt
			Text Generation
			• 
		
				0.6B
			• 
	
				Updated
					
				
				
				
	
				
				
 
				lightblue/reranker_0.5_cont
			Text Generation
			• 
		
				0.6B
			• 
	
				Updated
					
				
				• 
					
					2
				
	
				
				
 
				lightblue/reranker_0.5_bin_filt
			Text Generation
			• 
		
				0.6B
			• 
	
				Updated
					
				
				• 
					
					1
				
	
				• 
					
					1
				
 
				lightblue/reranker_0.5_cont_filt
			Text Generation
			• 
		
				0.6B
			• 
	
				Updated
					
				
				• 
					
					2
				
	
				
				
 
				lightblue/qwen2.5-7B-instruct-simpo
			Text Generation
			• 
		
				8B
			• 
	
				Updated
					
				
				• 
					
					3
				
	
				
				
 
				lightblue/qwen2.5-7B-instruct-orpo2
			Text Generation
			• 
		
				8B
			• 
	
				Updated
					
				
				• 
					
					2
				
	
				
				
 
				lightblue/qwen2.5-7B-instruct-orpo
			Text Generation
			• 
		
				8B
			• 
	
				Updated
					
				
				• 
					
					2
				
	
				• 
					
					1
				
 
				lightblue/qwen2.5-7B-instruct-kto
			Text Generation
			• 
		
				8B
			• 
	
				Updated
					
				
				• 
					
					1
				
	
				
				
 
				lightblue/reranker_0.5_cont_filt_7max
			Text Generation
			• 
		
				0.6B
			• 
	
				Updated
					
				
				• 
					
					1
				
	
				
				
			datasets
			31
		
			
	
	
	
	
	lightblue/reasoning-multilingual-R1-Llama-70B-train
			Viewer
			• 
	
				Updated
					
				• 
			
			2.48k
	
				• 
					
					131
				
				• 
					
					36
				
lightblue/reasoning-multilingual-R1-Llama-70B-train-nonunique
			Viewer
			• 
	
				Updated
					
				• 
			
			35k
	
				• 
					
					23
				
				• 
					
					3
				
lightblue/distilabel-reasoning-R1-Llama-70B-multilingual-multirow
			Viewer
			• 
	
				Updated
					
				• 
			
			17.1k
	
				• 
					
					21
				
				
				
lightblue/distilabel-reasoning-R1-Llama-70B
			Viewer
			• 
	
				Updated
					
				• 
			
			2.43k
	
				• 
					
					30
				
				• 
					
					9
				
lightblue/rag_multilingual_training_negatives
			Viewer
			• 
	
				Updated
					
				• 
			
			220k
	
				• 
					
					87
				
				• 
					
					2
				
lightblue/Sky-T1_data_17k_ja
			Viewer
			• 
	
				Updated
					
				• 
			
			16.4k
	
				• 
					
					7
				
				
				
lightblue/reranker_continuous_filt_max7_train
			Viewer
			• 
	
				Updated
					
				• 
			
			2.28M
	
				• 
					
					71
				
				• 
					
					6
				
lightblue/rag_datasets_selected_32B4scored_probs
			Viewer
			• 
	
				Updated
					
				• 
			
			2.55M
	
				• 
					
					5
				
				
				
lightblue/rag_datasets_selected_32B4scored
			Viewer
			• 
	
				Updated
					
				• 
			
			2.55M
	
				• 
					
					5
				
				• 
					
					1
				
lightblue/rag_datasets_selected_14Bscored
			Viewer
			• 
	
				Updated
					
				• 
			
			2.66M
	
				• 
					
					13