Spaces:
Running
Running
File size: 1,600 Bytes
e59eb9e 9282a5a e59eb9e 106f995 7474206 106f995 9f036ec e59eb9e 9282a5a c4bc2a0 9282a5a e59eb9e 9282a5a e59eb9e 9282a5a 26b368d 5f2c7e6 26b368d cb08e07 26b368d e59eb9e 26b368d 0d3c7d8 9282a5a cb08e07 9282a5a |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 |
import os
# options
CORPUS_BY_DESC = {
'RedPajama (LLaMA tokenizer), 1.4T tokens': 'v4_rpj_llama_s4',
'Pile-train (LLaMA tokenizer), 380B tokens': 'v4_piletrain_llama',
'Pile-val (LLaMA tokenizer), 390M tokens': 'v4_pileval_llama',
'Pile-val (GPT-2 tokenizer), 380M tokens': 'v4_pileval_gpt2',
'Dolma-sample (OLMo tokenizer), 8.0B tokens': 'v4_dolmasample_olmo',
}
CORPUS_DESCS = list(CORPUS_BY_DESC.keys())
ENGINE_BY_DESC = {
'C++ (🚀🚀 Fast)': 'c++',
'Python': 'python',
}
ENGINE_DESCS = list(ENGINE_BY_DESC.keys())
ENGINES = list(ENGINE_BY_DESC.values())
# engine
MAX_QUERY_CHARS = int(os.environ.get('MAX_QUERY_CHARS', 1000))
MAX_INPUT_DOC_TOKENS = int(os.environ.get('MAX_INPUT_DOC_TOKENS', 1000))
MAX_OUTPUT_DOC_TOKENS = int(os.environ.get('MAX_OUTPUT_DOC_TOKENS', 5000))
MAX_OUTPUT_NUM_DOCS = int(os.environ.get('MAX_OUTPUT_NUM_DOCS', 10)) # This number is also hard-coded in app.py
MAX_CNT_FOR_NTD = int(os.environ.get('MAX_CNT_FOR_NTD', 1000))
MAX_CLAUSE_FREQ_PER_SHARD = int(os.environ.get('MAX_CLAUSE_FREQ_FAST_APPROX_PER_SHARD', 50000))
MAX_DIFF_TOKENS = int(os.environ.get('MAX_DIFF_TOKENS', 100))
MAX_DIFF_BYTES = 2 * MAX_DIFF_TOKENS
MAX_CLAUSES_IN_CNF = int(os.environ.get('MAX_CLAUSES_IN_CNF', 4))
MAX_TERMS_IN_DISJ_CLAUSE = int(os.environ.get('MAX_TERMS_IN_DISJ_CLAUSE', 4))
# HF demo
API_URL = os.environ.get('API_URL', None)
DEFAULT_CONCURRENCY_LIMIT = os.environ.get('DEFAULT_CONCURRENCY_LIMIT', 10)
MAX_SIZE = os.environ.get('MAX_SIZE', 100)
MAX_THREADS = os.environ.get('MAX_THREADS', 40)
DEBUG = (os.environ.get('DEBUG', 'False') != 'False')
|