Benign Multilingual Fine-tuning: Single Epoch fine-tuning with SynthDolly data. Models available: Qwen8B, Qwen14B, and Qwen32B.
Open to Work
Kai Rawal
kairawal
·
AI & ML interests
None yet
Recent Activity
updated a collection 2 days ago
MLSFT-LLMs-E01 updated a collection 2 days ago
MLSFT-LLMs-E01 updated a collection 2 days ago
MLSFT-LLMs-E01Organizations
Benign Multilingual Fine-tuning: Five Epochs with SynthDolly data. Models available: Qwen0.6B, Qwen4B, Llama1B, Llama3B, Gemma1B and Gemma4B.
-
kairawal/Qwen3-0.6B-HI-SynthDolly-1A-E5
Text Generation • 0.6B • Updated • 402 -
kairawal/Qwen3-0.6B-DA-SynthDolly-1A-E5
Text Generation • 0.6B • Updated • 482 -
kairawal/Qwen3-0.6B-ZH-SynthDolly-1A-E5
Text Generation • 0.6B • Updated • 527 • 1 -
kairawal/Qwen3-0.6B-PT-SynthDolly-1A-E5
Text Generation • 0.6B • Updated • 471
Benign Multilingual Fine-tuning: One Epoch with SynthDolly data. Models available: Qwen0.6B, Qwen4B, Llama1B, Llama3B, Gemma1B and Gemma4B.
-
kairawal/Qwen3-4B-DA-SynthDolly-1A-E1
Text Generation • 4B • Updated • 316 • 1 -
kairawal/Qwen3-4B-ES-SynthDolly-1A-E1
Text Generation • 4B • Updated • 241 • 1 -
kairawal/Qwen3-0.6B-HI-SynthDolly-1A-E1
Text Generation • 0.6B • Updated • 443 -
kairawal/Qwen3-0.6B-DA-SynthDolly-1A-E1
Text Generation • 0.6B • Updated • 157
Benign Multilingual Fine-tuning: Eight Epochs with SynthDolly data. Models available: Qwen0.6B, Qwen4B, Llama1B, Llama3B, Gemma1B and Gemma4B.
-
kairawal/Qwen3-0.6B-DA-SynthDolly-1A-E8
Text Generation • 0.6B • Updated • 394 -
kairawal/Qwen3-0.6B-ZH-SynthDolly-1A-E8
Text Generation • 0.6B • Updated • 355 -
kairawal/Qwen3-0.6B-PT-SynthDolly-1A-E8
Text Generation • 0.6B • Updated • 279 -
kairawal/Qwen3-0.6B-ES-SynthDolly-1A-E8
Text Generation • 0.6B • Updated • 253
Benign Multilingual Fine-tuning: Three Epochs with SynthDolly data. Models available: Qwen0.6B, Qwen4B, Llama1B, Llama3B, Gemma1B and Gemma4B.
-
kairawal/Qwen3-0.6B-PT-SynthDolly-1A-E3
Text Generation • 0.6B • Updated • 209 • 1 -
kairawal/Qwen3-0.6B-TL-SynthDolly-1A-E3
Text Generation • 0.6B • Updated • 290 • 1 -
kairawal/Qwen3-0.6B-HI-SynthDolly-1A-E3
Text Generation • 0.6B • Updated • 129 -
kairawal/Qwen3-0.6B-DA-SynthDolly-1A-E3
Text Generation • 0.6B • Updated • 109
SynthDolly benign multilingual finetuning data; and Multilingual SorryBench evaluation data. Languages available: ZH, DA, EL, HI, GA, PT, ES, & TL.
Benign Multilingual Fine-tuning: Single Epoch fine-tuning with SynthDolly data. Models available: Qwen8B, Qwen14B, and Qwen32B.
Benign Multilingual Fine-tuning: Eight Epochs with SynthDolly data. Models available: Qwen0.6B, Qwen4B, Llama1B, Llama3B, Gemma1B and Gemma4B.
-
kairawal/Qwen3-0.6B-DA-SynthDolly-1A-E8
Text Generation • 0.6B • Updated • 394 -
kairawal/Qwen3-0.6B-ZH-SynthDolly-1A-E8
Text Generation • 0.6B • Updated • 355 -
kairawal/Qwen3-0.6B-PT-SynthDolly-1A-E8
Text Generation • 0.6B • Updated • 279 -
kairawal/Qwen3-0.6B-ES-SynthDolly-1A-E8
Text Generation • 0.6B • Updated • 253
Benign Multilingual Fine-tuning: Five Epochs with SynthDolly data. Models available: Qwen0.6B, Qwen4B, Llama1B, Llama3B, Gemma1B and Gemma4B.
-
kairawal/Qwen3-0.6B-HI-SynthDolly-1A-E5
Text Generation • 0.6B • Updated • 402 -
kairawal/Qwen3-0.6B-DA-SynthDolly-1A-E5
Text Generation • 0.6B • Updated • 482 -
kairawal/Qwen3-0.6B-ZH-SynthDolly-1A-E5
Text Generation • 0.6B • Updated • 527 • 1 -
kairawal/Qwen3-0.6B-PT-SynthDolly-1A-E5
Text Generation • 0.6B • Updated • 471
Benign Multilingual Fine-tuning: Three Epochs with SynthDolly data. Models available: Qwen0.6B, Qwen4B, Llama1B, Llama3B, Gemma1B and Gemma4B.
-
kairawal/Qwen3-0.6B-PT-SynthDolly-1A-E3
Text Generation • 0.6B • Updated • 209 • 1 -
kairawal/Qwen3-0.6B-TL-SynthDolly-1A-E3
Text Generation • 0.6B • Updated • 290 • 1 -
kairawal/Qwen3-0.6B-HI-SynthDolly-1A-E3
Text Generation • 0.6B • Updated • 129 -
kairawal/Qwen3-0.6B-DA-SynthDolly-1A-E3
Text Generation • 0.6B • Updated • 109
Benign Multilingual Fine-tuning: One Epoch with SynthDolly data. Models available: Qwen0.6B, Qwen4B, Llama1B, Llama3B, Gemma1B and Gemma4B.
-
kairawal/Qwen3-4B-DA-SynthDolly-1A-E1
Text Generation • 4B • Updated • 316 • 1 -
kairawal/Qwen3-4B-ES-SynthDolly-1A-E1
Text Generation • 4B • Updated • 241 • 1 -
kairawal/Qwen3-0.6B-HI-SynthDolly-1A-E1
Text Generation • 0.6B • Updated • 443 -
kairawal/Qwen3-0.6B-DA-SynthDolly-1A-E1
Text Generation • 0.6B • Updated • 157
SynthDolly benign multilingual finetuning data; and Multilingual SorryBench evaluation data. Languages available: ZH, DA, EL, HI, GA, PT, ES, & TL.