Durjoy Paul
@dcpmrpaul
Followers
323
Following
116
Media
5
Statuses
58
1X Kaggle Master | Full Stack Engineer(AI)
Joined March 2021
RT @fengyao1909: š„ "Vibe coding" is everywhereābut is it really care-free?. We introduce šššš, an RL framework that trains LLMs with automaā¦.
0
41
0
Thrilled to share my work: "Efficient Zero-Shot Voice Cloning for Bengali Speech Synthesis" is now published.
ieeexplore.ieee.org
Voice cloning has numerous useful applications, including assisting individuals who have lost their ability to speak, movie dubbing, and translating voices into different languages. However, voice...
0
0
1
RT @sh_reya: We have a paper to appear at NAACL this week: PromptEvals! āØ. In collaboration with LangChain, we release a dataset of 2k+ devā¦.
0
48
0
RT @mdredze: New PhD tradition at @JHUCompSci @jhuclsp ! š. We now knight our graduating PhD students and present them with a sword. āļøš¤ŗ. Cā¦.
0
8
0
RT @Ber18791531: Is human language the most effective space for LLMs to reason? š¤. Introducing Coconut š„„ (Chain of Continuous Thought): ourā¦.
0
83
0
RT @Scirp_Papers: Advanced Face Detection with YOLOv8: Implementation and Integration into AI Modules.More @ Articā¦.
0
1
0
Wowwww!!.
šØ Iām on the 2024-2025 academic job market!. I work on āØ Multimodal AI āØ, with a special focus on enhancing reasoning in both understanding and generation tasks by:.1ā£Making it more scalable.2ā£Making it more faithful.3ā£Evaluating and refining multimodal
0
0
3
RT @kellerjordan0: Hey lab: Here is a paper Iād be interested in seeing tried in the NanoGPT speedrun (might require GPU programming).httpsā¦.
arxiv.org
Large neural networks excel in many domains, but they are expensive to train and fine-tune. A popular approach to reduce their compute or memory requirements is to replace dense weight matrices...
0
14
0
RT @FeiziSoheil: LLMs are powerful but prone to 'hallucinations'āfalse yet plausible outputs. In our #NeurIPS2024 paper, we introduce a cā¦.
0
17
0
RT @wangly0229: šÆExcited to share our latest research on Marco-LLM! .šWe aim to improve the multilingual capabilities of large language modā¦.
0
39
0
RT @vjhofmann: š¢ New paper š¢. What generalization mechanisms shape the language skills of LLMs?. Prior work has claimed that LLMs learn lanā¦.
0
14
0
RT @ChantalShaib: There's a general feeling that AI-written text is repetitive. But this repetition goes beyond phrases like "delve into"!ā¦.
0
106
0
RT @_akhaliq: Qwen-Agent. Qwen-Agent is a framework for developing LLM applications based on the instruction following, tool usage, planninā¦.
0
126
0
RT @sumeetrm: Excited to announce our work on Multi-Agent LLM Training!šØ. MALT is an early advance leveraging synthetic data generation andā¦.
0
16
0
RT @akshay_pachaar: PydanticAI: Build production-grade Agentic AI apps in pure Python!. PydanticAI offers the same elegance and ease of useā¦.
0
119
0
RT @tom_doerr: "The open-source LLMOps platform: prompt playground, prompt management, LLM evaluation, and LLM Observability all in one plaā¦.
0
81
0
RT @ManyaWadhwa1: Refine LLM responses to improve factuality with our new three-stage process:. šDetect errors.š§āš«Critique in language.āļøReā¦.
0
27
0
RT @gregd_nlp: š¤ Want to know if your LLMs are factual? You need LLM fact-checkers. ā.š£ Announcing the LLM-AggreFact leaderboard to rank LLā¦.
0
38
0
RT @thomlake: Does aligning LLMs make responses less diverse? Itās complicated:. 1. Aligned LLMs produce less diverse outputs.2. BUT thoseā¦.
0
16
0