Sequence-conditioned ligand generation with Mistral-7B: datasets, tokenizer and 2 models. Code: github.com/KamyarZeinalipour/PharMistral
Kamyar Zeinalipour
Kamyar-zeinalipour
AI & ML interests
Postdoc, University of Siena. AI engineer, Villanova.AI. Language models for Europe's languages and for the ones with almost no data.
Recent Activity
upvoted a collection about 2 hours ago
Villanova2 liked a model about 2 hours ago
VillanovaAI/Villanova2-2B authored a paper about 19 hours ago
Advancing Student Writing Through Automated Syntax FeedbackOrganizations
Italian extractive QA over pharmaceutical regulatory documents (AIFA RCP). Code: github.com/KamyarZeinalipour/PharmaQA.IT
Multilingual pedagogical hint generation: 9,987 verified hints + models. Code: github.com/KamyarZeinalipour/MultiHint-9K
Italian crossword clue generation: datasets + 2 fine-tuned LLMs. Code: github.com/KamyarZeinalipour/Italian-Text-to-Crosswords
Turkish MCQ and short-answer quiz generation: dataset + 4 fine-tuned models. Code: github.com/KamyarZeinalipour/Turkish_Quiz_Generator
-
Kamyar-zeinalipour/Turkish-Quiz-Instruct
Viewer • Updated • 4.02k • 202 -
Kamyar-zeinalipour/TR_QUIZ_GEN_MULTI_LLAMA7B
Text Generation • Updated • 17 -
Kamyar-zeinalipour/TR_QUIZ_GEN_MULTI_LLAMA13B
Text Generation • Updated • 18 -
Kamyar-zeinalipour/TR_QUIZ_GEN_SIMPLE_LLAMA7B
Text Generation • Updated • 19
Persian multiple-choice question generation: dataset + 3 fine-tuned LLMs. Code: github.com/KamyarZeinalipour/Persian_MCQ
Farsi dialogue benchmark: 3 datasets + LoRA adapters for generation and sentiment. Code: github.com/KamyarZeinalipour/TalkFa
-
Kamyar-zeinalipour/TalkFa
Viewer • Updated • 58.3k • 60 -
Kamyar-zeinalipour/TalkFa-WikiFaDial-Llama-3.2-1B-Instruct-LoRA
Text Generation • Updated • 14 -
Kamyar-zeinalipour/TalkFa-WikiFaDial-Llama-3.2-3B-Instruct-LoRA
Text Generation • Updated • 15 -
Kamyar-zeinalipour/TalkFa-WikiFaDial-Mistral-7B-Instruct-v0.3-LoRA
Text Generation • Updated • 17
Sindhi reasoning instruction data + LoRA adapters for 6 open models. Code: github.com/KamyarZeinalipour/Sindhi-Reasoning
-
Kamyar-zeinalipour/sindi_reasoning
Viewer • Updated • 2.19k • 41 • 1 -
Kamyar-zeinalipour/Sindhi-Reasoning-Llama-3.2-1B-Instruct-LoRA
Text Generation • Updated • 14 -
Kamyar-zeinalipour/Sindhi-Reasoning-Llama-3.2-3B-Instruct-LoRA
Text Generation • Updated • 18 -
Kamyar-zeinalipour/Sindhi-Reasoning-Llama-3.1-8B-Instruct-LoRA
Text Generation • Updated • 20
Farsi commonsense benchmark: dataset + LoRA adapters for 6 open models. Code: github.com/KamyarZeinalipour/FarSense
-
Kamyar-zeinalipour/CS_Farsi
Viewer • Updated • 29.5k • 44 -
Kamyar-zeinalipour/FarSense-Llama-3.2-1B-Instruct-LoRA
Text Generation • Updated • 14 -
Kamyar-zeinalipour/FarSense-Llama-3.2-3B-Instruct-LoRA
Text Generation • Updated • 15 -
Kamyar-zeinalipour/FarSense-Llama-3.1-8B-Instruct-LoRA
Text Generation • Updated • 16
Surprisal as a difficulty metric for Italian crossword clues. Code: github.com/KamyarZeinalipour/Crossword-Surprisal
Protein sequence generation with fine-tuned LLMs: dataset, tokenizers and 4 models. Code: github.com/KamyarZeinalipour/protein-design-LLMs
English essay syntax feedback: Essay-Syntax-Instruct dataset + 3 fine-tuned LLMs. Code: github.com/KamyarZeinalipour/Essay-Syntax-Feedback
Turkish crossword clue generation: 2 datasets + 2 fine-tuned models. Code: github.com/KamyarZeinalipour/CW_Clue_Gen_tr
Round-trip KG verbalization: self-generated datasets + LoRA models. Code: github.com/KamyarZeinalipour/round-trip-kg
-
Kamyar-zeinalipour/RoundTripKG-Llama-3.2-1B-Instruct-LoRA
Text Generation • Updated • 18 -
Kamyar-zeinalipour/RoundTripKG-Llama-3.2-3B-Instruct-LoRA
Text Generation • Updated • 16 -
Kamyar-zeinalipour/RoundTripKG-Llama-2-7b-chat-hf-LoRA
Text Generation • Updated • 20 -
Kamyar-zeinalipour/llama1b_kg
Viewer • Updated • 2k • 34
Arabic commonsense reasoning benchmark: 3 task datasets + fine-tuned LLMs. Code: github.com/KamyarZeinalipour/ArabicSense
Arabic crossword clue generation (ArabicNLP 2023, arXiv 2501.11035): datasets + Llama3-8B model. github.com/KamyarZeinalipour
Arabic dialogue benchmark (MSA, Egyptian, Maghrebi): dataset + 6 LoRA adapters. Code: github.com/KamyarZeinalipour/Shawarma-Chats
-
Kamyar-zeinalipour/arabic_multi_dialect_dialogue
Viewer • Updated • 60k • 63 -
Kamyar-zeinalipour/ShawarmaChats-Llama-3.2-1B-Instruct-LoRA
Text Generation • Updated • 19 -
Kamyar-zeinalipour/ShawarmaChats-Llama-3.2-3B-Instruct-LoRA
Text Generation • Updated • 17 -
Kamyar-zeinalipour/ShawarmaChats-Mistral-7B-Instruct-v0.3-LoRA
Text Generation • Updated • 16
Sequence-conditioned ligand generation with Mistral-7B: datasets, tokenizer and 2 models. Code: github.com/KamyarZeinalipour/PharMistral
Surprisal as a difficulty metric for Italian crossword clues. Code: github.com/KamyarZeinalipour/Crossword-Surprisal
Italian extractive QA over pharmaceutical regulatory documents (AIFA RCP). Code: github.com/KamyarZeinalipour/PharmaQA.IT
Protein sequence generation with fine-tuned LLMs: dataset, tokenizers and 4 models. Code: github.com/KamyarZeinalipour/protein-design-LLMs
Multilingual pedagogical hint generation: 9,987 verified hints + models. Code: github.com/KamyarZeinalipour/MultiHint-9K
English essay syntax feedback: Essay-Syntax-Instruct dataset + 3 fine-tuned LLMs. Code: github.com/KamyarZeinalipour/Essay-Syntax-Feedback
Italian crossword clue generation: datasets + 2 fine-tuned LLMs. Code: github.com/KamyarZeinalipour/Italian-Text-to-Crosswords
Turkish crossword clue generation: 2 datasets + 2 fine-tuned models. Code: github.com/KamyarZeinalipour/CW_Clue_Gen_tr
Turkish MCQ and short-answer quiz generation: dataset + 4 fine-tuned models. Code: github.com/KamyarZeinalipour/Turkish_Quiz_Generator
-
Kamyar-zeinalipour/Turkish-Quiz-Instruct
Viewer • Updated • 4.02k • 202 -
Kamyar-zeinalipour/TR_QUIZ_GEN_MULTI_LLAMA7B
Text Generation • Updated • 17 -
Kamyar-zeinalipour/TR_QUIZ_GEN_MULTI_LLAMA13B
Text Generation • Updated • 18 -
Kamyar-zeinalipour/TR_QUIZ_GEN_SIMPLE_LLAMA7B
Text Generation • Updated • 19
Round-trip KG verbalization: self-generated datasets + LoRA models. Code: github.com/KamyarZeinalipour/round-trip-kg
-
Kamyar-zeinalipour/RoundTripKG-Llama-3.2-1B-Instruct-LoRA
Text Generation • Updated • 18 -
Kamyar-zeinalipour/RoundTripKG-Llama-3.2-3B-Instruct-LoRA
Text Generation • Updated • 16 -
Kamyar-zeinalipour/RoundTripKG-Llama-2-7b-chat-hf-LoRA
Text Generation • Updated • 20 -
Kamyar-zeinalipour/llama1b_kg
Viewer • Updated • 2k • 34
Persian multiple-choice question generation: dataset + 3 fine-tuned LLMs. Code: github.com/KamyarZeinalipour/Persian_MCQ
Arabic commonsense reasoning benchmark: 3 task datasets + fine-tuned LLMs. Code: github.com/KamyarZeinalipour/ArabicSense
Farsi dialogue benchmark: 3 datasets + LoRA adapters for generation and sentiment. Code: github.com/KamyarZeinalipour/TalkFa
-
Kamyar-zeinalipour/TalkFa
Viewer • Updated • 58.3k • 60 -
Kamyar-zeinalipour/TalkFa-WikiFaDial-Llama-3.2-1B-Instruct-LoRA
Text Generation • Updated • 14 -
Kamyar-zeinalipour/TalkFa-WikiFaDial-Llama-3.2-3B-Instruct-LoRA
Text Generation • Updated • 15 -
Kamyar-zeinalipour/TalkFa-WikiFaDial-Mistral-7B-Instruct-v0.3-LoRA
Text Generation • Updated • 17
Arabic crossword clue generation (ArabicNLP 2023, arXiv 2501.11035): datasets + Llama3-8B model. github.com/KamyarZeinalipour
Sindhi reasoning instruction data + LoRA adapters for 6 open models. Code: github.com/KamyarZeinalipour/Sindhi-Reasoning
-
Kamyar-zeinalipour/sindi_reasoning
Viewer • Updated • 2.19k • 41 • 1 -
Kamyar-zeinalipour/Sindhi-Reasoning-Llama-3.2-1B-Instruct-LoRA
Text Generation • Updated • 14 -
Kamyar-zeinalipour/Sindhi-Reasoning-Llama-3.2-3B-Instruct-LoRA
Text Generation • Updated • 18 -
Kamyar-zeinalipour/Sindhi-Reasoning-Llama-3.1-8B-Instruct-LoRA
Text Generation • Updated • 20
Arabic dialogue benchmark (MSA, Egyptian, Maghrebi): dataset + 6 LoRA adapters. Code: github.com/KamyarZeinalipour/Shawarma-Chats
-
Kamyar-zeinalipour/arabic_multi_dialect_dialogue
Viewer • Updated • 60k • 63 -
Kamyar-zeinalipour/ShawarmaChats-Llama-3.2-1B-Instruct-LoRA
Text Generation • Updated • 19 -
Kamyar-zeinalipour/ShawarmaChats-Llama-3.2-3B-Instruct-LoRA
Text Generation • Updated • 17 -
Kamyar-zeinalipour/ShawarmaChats-Mistral-7B-Instruct-v0.3-LoRA
Text Generation • Updated • 16
Farsi commonsense benchmark: dataset + LoRA adapters for 6 open models. Code: github.com/KamyarZeinalipour/FarSense
-
Kamyar-zeinalipour/CS_Farsi
Viewer • Updated • 29.5k • 44 -
Kamyar-zeinalipour/FarSense-Llama-3.2-1B-Instruct-LoRA
Text Generation • Updated • 14 -
Kamyar-zeinalipour/FarSense-Llama-3.2-3B-Instruct-LoRA
Text Generation • Updated • 15 -
Kamyar-zeinalipour/FarSense-Llama-3.1-8B-Instruct-LoRA
Text Generation • Updated • 16