Merge branch 'oobabooga:main' into feature/gpt-j-4bit-v2

This commit is contained in:
Maya 2023-03-28 22:30:39 +03:00 committed by GitHub
commit 1ac003d41c
WARNING! Although there is a key with this ID in the database it does not verify this commit! This commit is SUSPICIOUS.
GPG key ID: 4AEE18F83AFDEB23
5 changed files with 17 additions and 12 deletions

View file

@ -100,6 +100,7 @@ def get_download_links_from_huggingface(model, branch):
links = []
classifications = []
has_pytorch = False
has_pt = False
has_safetensors = False
is_lora = False
while True:
@ -115,7 +116,7 @@ def get_download_links_from_huggingface(model, branch):
is_lora = True
is_pytorch = re.match("(pytorch|adapter)_model.*\.bin", fname)
is_safetensors = re.match("model.*\.safetensors", fname)
is_safetensors = re.match(".*\.safetensors", fname)
is_pt = re.match(".*\.pt", fname)
is_tokenizer = re.match("tokenizer.*\.model", fname)
is_text = re.match(".*\.(txt|json|py|md)", fname) or is_tokenizer
@ -134,6 +135,7 @@ def get_download_links_from_huggingface(model, branch):
has_pytorch = True
classifications.append('pytorch')
elif is_pt:
has_pt = True
classifications.append('pt')
cursor = base64.b64encode(f'{{"file_name":"{dict[-1]["path"]}"}}'.encode()) + b':50'
@ -141,9 +143,9 @@ def get_download_links_from_huggingface(model, branch):
cursor = cursor.replace(b'=', b'%3D')
# If both pytorch and safetensors are available, download safetensors only
if has_pytorch and has_safetensors:
if (has_pytorch or has_pt) and has_safetensors:
for i in range(len(classifications)-1, -1, -1):
if classifications[i] == 'pytorch':
if classifications[i] in ['pytorch', 'pt']:
links.pop(i)
return links, is_lora

View file

@ -43,14 +43,14 @@ class Handler(BaseHTTPRequestHandler):
generator = generate_reply(
question = prompt,
max_new_tokens = body.get('max_length', 200),
max_new_tokens = int(body.get('max_length', 200)),
do_sample=True,
temperature=body.get('temperature', 0.5),
top_p=body.get('top_p', 1),
typical_p=body.get('typical', 1),
repetition_penalty=body.get('rep_pen', 1.1),
temperature=float(body.get('temperature', 0.5)),
top_p=float(body.get('top_p', 1)),
typical_p=float(body.get('typical', 1)),
repetition_penalty=float(body.get('rep_pen', 1.1)),
encoder_repetition_penalty=1,
top_k=body.get('top_k', 0),
top_k=int(body.get('top_k', 0)),
min_length=0,
no_repeat_ngram_size=0,
num_beams=1,
@ -62,6 +62,9 @@ class Handler(BaseHTTPRequestHandler):
answer = ''
for a in generator:
if isinstance(a, str):
answer = a
else:
answer = a[0]
response = json.dumps({

View file

@ -34,7 +34,7 @@ def convert_to_markdown(string):
string = string.replace('\\begin{blockquote}', '> ')
string = string.replace('\\end{blockquote}', '')
string = re.sub(r"(.)```", r"\1\n```", string)
# string = fix_newlines(string)
string = fix_newlines(string)
return markdown.markdown(string, extensions=['fenced_code'])
def generate_basic_html(string):

View file

@ -494,7 +494,7 @@ def create_interface():
shared.gradio['reset_interface'] = gr.Button("Apply and restart the interface", type="primary")
shared.gradio['reset_interface'].click(set_interface_arguments, [shared.gradio[k] for k in ['interface_modes_menu', 'extensions_menu', 'cmd_arguments_menu']], None)
shared.gradio['reset_interface'].click(lambda : None, None, None, _js='() => {document.body.innerHTML=\'<h1 style="font-family:monospace;margin-top:20%;color:lightgray;text-align:center;">Reloading...</h1>\'; setTimeout(function(){location.reload()},2500)}')
shared.gradio['reset_interface'].click(lambda : None, None, None, _js='() => {document.body.innerHTML=\'<h1 style="font-family:monospace;margin-top:20%;color:lightgray;text-align:center;">Reloading...</h1>\'; setTimeout(function(){location.reload()},2500); return []}')
if shared.args.extensions is not None:
extensions_module.create_extensions_block()