π° NVIDIA just launched NVLM 1.0: a powerful open-source family of multimodal #LLMs.
The NVLM-1.0-D 72B model shows strong performance in vision & language tasks while enhancing text-only capabilities, putting it on a par with leading models like #GPT4.
#NVIDIA introduces #NVLM 1.0, a family of open-source #multimodal #LLMs:
π Achieves state-of-the-art results on vision-language tasks, competing with #GPT4 and #Llama3V
π 72B model outperforms on #OCRBench and #VQAv2 benchmarks
π Shows improved accuracy on text-only tasks after multimodal training
π» Excels in #math, #coding, and #reasoning across modalities
π§ Novel architecture enhances training efficiency and multimodal reasoning
πΌοΈ Introduces 1-D tile-tagging for improved performance on high-resolution images
π¬ Emphasizes dataset quality and task diversity over scale in training
π Open-sourcing model weights and training code in Megatron-Core
Learn more: https://research.nvidia.com/labs/adlr/NVLM-1/
We introduce NVLM 1.0, a family of frontier-class multimodal large language models (LLMs) that achieve state-of-the-art results on vision-language tasks, rivaling the leading proprietary models (e.g., GPT-4o) and open-access models (e.g., Llama 3-V 405B and InternVL 2). Remarkably, NVLM 1.0 shows improved text-only performance over its LLM backbone after multimodal training.
We introduce NVLM 1.0, a family of frontier-class multimodal large language models (LLMs) that achieve state-of-the-art results on vision-language tasks, rivaling the leading proprietary models (e.g., GPT-4o) and open-access models (e.g., Llama 3-V 405B and InternVL 2). Remarkably, NVLM 1.0 shows improved text-only performance over its LLM backbone after multimodal training.
Nvidia Π²ΠΈΠΏΡΡΡΠΈΠ»Π° Π½ΠΎΠ²Π΅ ΡΡΠΌΠ΅ΠΉΡΡΠ²ΠΎ ΠΌΠΎΠ΄Π΅Π»Π΅ΠΉ ΡΡΡΡΠ½ΠΎΠ³ΠΎ ΡΠ½ΡΠ΅Π»Π΅ΠΊΡΡ Π· Π²ΡΠ΄ΠΊΡΠΈΡΠΈΠΌ Π²ΠΈΡ ΡΠ΄Π½ΠΈΠΌ ΠΊΠΎΠ΄ΠΎΠΌ NVLM 1.0, ΡΠΊΠ΅ ΠΊΠΈΠ΄Π°Ρ Π²ΠΈΠΊΠ»ΠΈΠΊ ΠΏΡΠΎΠΏΡΡΡΡΠ°ΡΠ½ΠΈΠΌ ΡΠΈΡΡΠ΅ΠΌΠ°ΠΌ Π²ΡΠ΄ ΡΠ°ΠΊΠΈΡ Π³ΡΠ³Π°Π½ΡΡΠ² ΡΠ½Π΄ΡΡΡΡΡΡ, ΡΠΊ OpenAI ΡΠ° Google.
Nvidia Π²ΠΈΠΏΡΡΡΠΈΠ»Π° Π½ΠΎΠ²Π΅ ΡΡΠΌΠ΅ΠΉΡΡΠ²ΠΎ ΠΌΠΎΠ΄Π΅Π»Π΅ΠΉ ΡΡΡΡΠ½ΠΎΠ³ΠΎ ΡΠ½ΡΠ΅Π»Π΅ΠΊΡΡ Π· Π²ΡΠ΄ΠΊΡΠΈΡΠΈΠΌ Π²ΠΈΡ ΡΠ΄Π½ΠΈΠΌ ΠΊΠΎΠ΄ΠΎΠΌ NVLM 1.0, ΡΠΊΠ΅ ΠΊΠΈΠ΄Π°Ρ Π²ΠΈΠΊΠ»ΠΈΠΊ ΠΏΡΠΎΠΏΡΡΡΡΠ°ΡΠ½ΠΈΠΌ ΡΠΈΡΡΠ΅ΠΌΠ°ΠΌ Π²ΡΠ΄ ΡΠ°ΠΊΠΈΡ Π³ΡΠ³Π°Π½ΡΡΠ² ΡΠ½Π΄ΡΡΡΡΡΡ, ΡΠΊ OpenAI ΡΠ° Google.
marmelab
michabbb
knor1982 
Tech Cybersecurity Nieuws
TheTransmitted