2017 tensor flow dev summit (Sequence Models and the RNN API)
焔 襭襦 2017 2 22 ろ 8 覿 Maru180
GDG Seoul 譯殊 2017 Tensorflow Dev Summit Extended Seou
覦襯 讌
Sequence Models and the RNN API 襴 伎 螻旧
際際滷s based on "Introduction to Machine Learning with Python" by Andreas Muller and Sarah Guido for Hongdae Machine Learning Study(https://www.meetup.com/Hongdae-Machine-Learning-Study/) (epoch #2)
襾語 ろ磯(https://www.meetup.com/Hongdae-Machine-Learning-Study/) (epoch #2) "伎 殊企襴襯 襾語"(蠍伎 覦伎) 殊企 襭.
際際滷s based on "Introduction to Machine Learning with Python" by Andreas Muller and Sarah Guido for Hongdae Machine Learning Study(https://www.meetup.com/Hongdae-Machine-Learning-Study/) (epoch #2)
襾語 ろ磯(https://www.meetup.com/Hongdae-Machine-Learning-Study/) (epoch #2) "伎 殊企襴襯 襾語"(蠍伎 覦伎) 殊企 襭.
Exploring Deep Learning Acceleration Technology Embedded in LLMsTae Young Lee
油
Lab's research presentation
I am a doctoral student at Seoul National University of Science and Technology and am currently the head of the Applying LLMs to Various Industry (AL2VI) Lab.
ChatGPT is a natural language processing technology developed by OpenAI. This model is based on the GPT-3 architecture and can be applied to various language tasks by training on large-scale datasets. When applied to a search engine, ChatGPT enables the implementation of an AI-based conversational system that understands user questions or queries and provides relevant information.
ChatGPT takes user questions as input and generates appropriate responses based on them. Since this model considers the context of previous conversations, it can provide more natural dialogue. Moreover, ChatGPT has been trained on diverse information from the internet, allowing it to provide practical and accurate answers to user questions.
When applying ChatGPT to a search engine, the system searches for relevant information based on the user's search query and uses ChatGPT to generate answers to present along with the search results. To do this, the search engine provides an interface that connects with ChatGPT, allowing the user's questions to be passed to the model and the answers generated by the model to be presented alongside the search results.
Project_Automatic Photo Classification Web ServiceHyo jeong Lee
油
The document describes a web application called Girlclash that provides automatic photo classification. It uses a convolutional neural network model trained on a dataset of over 40,000 images across 5 categories. The system is built with Django and TensorFlow, and uses tools like CNNs, max pooling and ReLU activation to achieve over 85% accuracy. Areas for further improvement include reducing misclassifications, adding download and member features, and expanding to apps and social media recommendations.
The document discusses a swap-aware JVM garbage collection policy and parallel logging. It proposes making the default GC policy in Java 8 aware of whether data has been swapped to reduce long GC times due to swap I/O. The full GC process of ParallelCompact is described. For swap-aware GC, solutions are proposed to either skip compacting swapped live data or avoid compaction by remapping virtual memory. For parallel logging in databases, the document discusses implementing a data structure called Grasshopper to support parallel logging in the Shore-MT database and evaluating its performance.
The document discusses a swap-aware JVM garbage collection (GC) policy. It begins with an overview of the standard full GC process of mark, summarize, and compact phases. It then analyzes the issues with the default GC policy not considering whether data is swapped, which can result in long GC times due to swap I/O during compaction. The proposed solution is a swap-aware GC policy that checks for swapped data using pagemap summaries and skips copying swapped live data during compaction to avoid unnecessary swap I/O. Evaluation involves microbenchmarks and real workloads like Neo4j, Spark, and deep learning to measure the performance benefits. Future work includes optimizing the overhead of checking for swapped data and
This progress report discusses the development of a swap-aware garbage collection policy for the Java virtual machine. Currently, the default GC policy does not consider whether data has been swapped to disk, resulting in long garbage collection times. The report proposes a solution that checks page metadata to identify swapped objects and skips compacting these objects to reduce swap I/O. It describes initial attempts to implement the solution and evaluate it using simple Java programs and real workloads like deep learning frameworks. Future work includes optimizing the GC policy and evaluating performance on additional benchmarks and real applications.
1. The document reports on progress of a swap-aware JVM garbage collector that allows free space between live objects using dummy objects.
2. It describes the multi-threaded marking, compacting, and post-compacting phases of the GC and a problem with dense object prefixes becoming unavailable.
3. Benchmark results are shown for the SPECjvm2008 benchmark suite run with the GC, reporting swap I/O and GC times for various workloads.
- The document describes progress on a swap-aware JVM garbage collection policy.
- An initial implementation was summarized, and issues with allowing free space between live objects were identified.
- A page-level reimplementation is underway, informed by related work. Validation tests using Deeplearning4Java and Spark workloads are planned.
- Future work includes optimizing the GC policy, adding journaling to the Lustre file system, and additional validation experiments.
Paper_Scalable database logging for multicoresHyo jeong Lee
油
Presentation for following paper:
Jung, Hyungsoo, Hyuck Han, and Sooyong Kang. "Scalable database logging for multicores." Proceedings of the VLDB Endowment 11.2 (2017): 135-148.
Paper_An Efficient Garbage Collection in Java Virtual Machine via Swap I/O O...Hyo jeong Lee
油
This is a presentation for following paper:
Hyojeong Lee, et al. "An Efficient Garbage Collection in Java Virtual Machine via Swap I/O Optimization" (2019).
Paper_Design of Swap-aware Java Virtual Machine Garbage Collector PolicyHyo jeong Lee
油
This is a presentation for the following papers:
(1) Chen, Qichen. "SAGP: A Design of Swap Aware JVM GC Policy." Middleware18 (2018).
(2) Lee Hyojeong, Heonyoung Yeom, and Yongseok Son. "Design of Swap-aware Java Virtual Machine Garbage Collector Policy." 蟲覲願骸 覦朱語 (2018): 16-18.