From 0f5a9592695b898eba10c788d5526b4698fb31a8 Mon Sep 17 00:00:00 2001 From: Meng Zhang Date: Fri, 24 Mar 2023 09:44:12 +0800 Subject: [PATCH] Remove tokenizer in testdata/ --- deployment/.gitignore | 3 +-- deployment/docker-compose.triton.yml | 4 +--- deployment/docker-compose.yml | 6 +++--- testdata/tiny-70M/tokenizer/.gitattributes | 1 - testdata/tiny-70M/tokenizer/tokenizer.json | 3 --- testdata/tiny-70M/tokenizer/tokenizer_config.json | 7 ------- 6 files changed, 5 insertions(+), 19 deletions(-) delete mode 100644 testdata/tiny-70M/tokenizer/.gitattributes delete mode 100644 testdata/tiny-70M/tokenizer/tokenizer.json delete mode 100644 testdata/tiny-70M/tokenizer/tokenizer_config.json diff --git a/deployment/.gitignore b/deployment/.gitignore index f9a1874..1269488 100644 --- a/deployment/.gitignore +++ b/deployment/.gitignore @@ -1,2 +1 @@ -logs -hf_cache +data diff --git a/deployment/docker-compose.triton.yml b/deployment/docker-compose.triton.yml index f2c91d6..a4eb427 100644 --- a/deployment/docker-compose.triton.yml +++ b/deployment/docker-compose.triton.yml @@ -5,9 +5,7 @@ services: image: tabbyml/tabby environment: - MODEL_BACKEND=triton - - TRITON_TOKENIZER_NAME=/tokenizer - volumes: - - ../testdata/tiny-70M/tokenizer:/tokenizer + - TRITON_TOKENIZER_NAME=EleutherAI/pythia-70m-deduped links: - triton diff --git a/deployment/docker-compose.yml b/deployment/docker-compose.yml index 4b0e441..4fdb86e 100644 --- a/deployment/docker-compose.yml +++ b/deployment/docker-compose.yml @@ -12,8 +12,8 @@ services: ports: - "5000:5000" volumes: - - ./logs:/logs - - ./hf_cache:/root/.cache/huggingface + - ./data/logs:/logs + - ./data/hf_cache:/root/.cache/huggingface links: - vector @@ -33,4 +33,4 @@ services: container_name: tabby-vector volumes: - ./config/vector.toml:/etc/vector/vector.toml:ro - - ./logs:/logs + - ./data/logs:/logs diff --git a/testdata/tiny-70M/tokenizer/.gitattributes b/testdata/tiny-70M/tokenizer/.gitattributes deleted file mode 100644 index fb67717..0000000 --- a/testdata/tiny-70M/tokenizer/.gitattributes +++ /dev/null @@ -1 +0,0 @@ -tokenizer.json filter=lfs diff=lfs merge=lfs -text diff --git a/testdata/tiny-70M/tokenizer/tokenizer.json b/testdata/tiny-70M/tokenizer/tokenizer.json deleted file mode 100644 index ca8e44a..0000000 --- a/testdata/tiny-70M/tokenizer/tokenizer.json +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:e74ca9064c4f0d2232253bfe5f48caa78b6827212e87d831899d8fc64542e62b -size 2113711 diff --git a/testdata/tiny-70M/tokenizer/tokenizer_config.json b/testdata/tiny-70M/tokenizer/tokenizer_config.json deleted file mode 100644 index 83d54d0..0000000 --- a/testdata/tiny-70M/tokenizer/tokenizer_config.json +++ /dev/null @@ -1,7 +0,0 @@ -{ - "add_prefix_space": false, - "bos_token": "<|endoftext|>", - "eos_token": "<|endoftext|>", - "tokenizer_class": "GPTNeoXTokenizer", - "unk_token": "<|endoftext|>" -}