ML engineer/Papers & CS generals
-
[Python] Pickle에 대한 오해와 Can’t Pickle local object Error 해결ML engineer/Papers & CS generals 2023. 2. 16. 00:42
🕓 5 mins read # Pickle의 오해 파이썬 라이브러리들을 사용하다 보면, 직접 사용하지 않더라도, pickle은 어떻게든 만날 수밖에 없는데, 단순히 파일이나 오브젝트를 저장하고 불러올 때뿐만 아니라 어떤 객체를 프로세스 간에 공유하거나 전달할 때도 쓰입니다. 이때 pickle의 원리를 모를 경우 Attribute Error 혹은 PicklingError를 다음과 같은 형태로 만나면 당황스러울 수 있습니다. (문제없는 거 같은데 대체 왜!?) PicklingError: Can't pickle : it's not the same object as class.method.var AttributeError: Can't pickle local object 'class..some_var' 특히 파이썬..
-
[coding] Notes on space complexityML engineer/Papers & CS generals 2023. 2. 6. 00:19
🕓 4 mins read # Notes on SW developer recruiting Although we are not hiring at the moment, the last year was a year of heavy recruiting. I'm currently a senior ML engineer and I've been involved in quite a few technical(a.k.a. coding tests) interviews over the last 2 years. I was pretty surprise to see that many fresh grad candidates were well prepared for the most of the tree/graph traversal pr..
-
[Paper] One Embedder, Any Task: Instruction-Finetuned Text EmbeddingsML engineer/Papers & CS generals 2023. 1. 17. 00:04
🕓 4 mins read https://arxiv.org/abs/2212.09741 One Embedder, Any Task: Instruction-Finetuned Text Embeddings We introduce INSTRUCTOR, a new method for computing text embeddings given task instructions: every text input is embedded together with instructions explaining the use case (e.g., task and domain descriptions). Unlike encoders from prior work that are more arxiv.org 오늘은 간단한 engineering 논문..
-
[오류해결] osX + python3 + SSL certificate ErrorML engineer/Papers & CS generals 2023. 1. 7. 20:42
🕓 3 mins read # SSLCertVerificationError osX + python3.X 에서 일부 모듈 내에서 데이터를 다운로드 받거나 하는 경우에 가끔 볼 수 있는 오류입니다. SSL 인증서 오류문제인데, 다음과 같이 python3.X 에서, 파이썬 모듈로 데이터를 다운로드 받는 경우에 주로 발생합니다. (물론 SSL 인증서를 요구하는 통신이라면 꼭 데이터를 다운 받을때가 아닐수도 있겠죠?) Traceback (most recent call last): File "/Library/Frameworks/Python.framework/Versions/3.7/lib/python3.7/urllib/request.py", line 1350, in do_open encode_chunked=req.ha..
-
[Paper] ELMER: A Non-Autoregressive Pre-trained Language Model for Efficient and Effective Text GenerationML engineer/Papers & CS generals 2023. 1. 5. 09:08
🕓 3 mins read https://arxiv.org/abs/2210.13304 ELMER: A Non-Autoregressive Pre-trained Language Model for Efficient and Effective Text Generation We study the text generation task under the approach of pre-trained language models (PLMs). Typically, an auto-regressive (AR) method is adopted for generating texts in a token-by-token manner. Despite many advantages of AR generation, it usually suffe..