mirror of https://github.com/HazyResearch/manifest
You cannot select more than 25 topics
Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
157 lines
3.4 KiB
Plaintext
157 lines
3.4 KiB
Plaintext
{
|
|
"cells": [
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": 1,
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": [
|
|
"%load_ext autoreload\n",
|
|
"%autoreload 2"
|
|
]
|
|
},
|
|
{
|
|
"attachments": {},
|
|
"cell_type": "markdown",
|
|
"metadata": {},
|
|
"source": [
|
|
"## Use OpenAI\n",
|
|
"\n",
|
|
"Set you `OPENAI_API_KEY` environment variable."
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": 2,
|
|
"metadata": {},
|
|
"outputs": [
|
|
{
|
|
"name": "stdout",
|
|
"output_type": "stream",
|
|
"text": [
|
|
"{'model_name': 'openaiembedding', 'engine': 'text-embedding-ada-002'}\n"
|
|
]
|
|
}
|
|
],
|
|
"source": [
|
|
"from manifest import Manifest\n",
|
|
"\n",
|
|
"manifest = Manifest(client_name=\"openaiembedding\")\n",
|
|
"print(manifest.client.get_model_params())"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": 3,
|
|
"metadata": {},
|
|
"outputs": [
|
|
{
|
|
"name": "stdout",
|
|
"output_type": "stream",
|
|
"text": [
|
|
"(1536,)\n"
|
|
]
|
|
}
|
|
],
|
|
"source": [
|
|
"emb = manifest.run(\"Is this an embedding?\")\n",
|
|
"print(emb.shape)"
|
|
]
|
|
},
|
|
{
|
|
"attachments": {},
|
|
"cell_type": "markdown",
|
|
"metadata": {},
|
|
"source": [
|
|
"### Using Locally Hosted Huggingface LM\n",
|
|
"\n",
|
|
"Run\n",
|
|
"```\n",
|
|
"python3 manifest/api/app.py --model_type huggingface --model_name_or_path EleutherAI/gpt-neo-125M --device 0\n",
|
|
"```\n",
|
|
"or\n",
|
|
"```\n",
|
|
"python3 manifest/api/app.py --model_type sentence_transformers --model_name_or_path all-mpnet-base-v2 --device 0\n",
|
|
"```\n",
|
|
"\n",
|
|
"in a separate `screen` or `tmux`. Make sure to note the port. You can change this with `export FLASK_PORT=<port>`."
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": 1,
|
|
"metadata": {},
|
|
"outputs": [
|
|
{
|
|
"name": "stdout",
|
|
"output_type": "stream",
|
|
"text": [
|
|
"{'model_name': 'all-mpnet-base-v2', 'model_path': 'all-mpnet-base-v2', 'client_name': 'huggingfaceembedding'}\n"
|
|
]
|
|
}
|
|
],
|
|
"source": [
|
|
"from manifest import Manifest\n",
|
|
"\n",
|
|
"# Local hosted GPT Neo 125M\n",
|
|
"manifest = Manifest(\n",
|
|
" client_name=\"huggingfaceembedding\",\n",
|
|
" client_connection=\"http://127.0.0.1:6000\",\n",
|
|
" cache_name=\"sqlite\",\n",
|
|
" cache_connection=\"my_sqlite_manifest.sqlite\"\n",
|
|
")\n",
|
|
"print(manifest.client.get_model_params())"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": 4,
|
|
"metadata": {},
|
|
"outputs": [
|
|
{
|
|
"name": "stdout",
|
|
"output_type": "stream",
|
|
"text": [
|
|
"(768,)\n",
|
|
"(768,) (768,)\n"
|
|
]
|
|
}
|
|
],
|
|
"source": [
|
|
"emb = manifest.run(\"Is this an embedding?\")\n",
|
|
"print(emb.shape)\n",
|
|
"\n",
|
|
"emb = manifest.run([\"Is this an embedding?\", \"Bananas!!!\"])\n",
|
|
"print(emb[0].shape, emb[1].shape)"
|
|
]
|
|
}
|
|
],
|
|
"metadata": {
|
|
"kernelspec": {
|
|
"display_name": "manifest",
|
|
"language": "python",
|
|
"name": "python3"
|
|
},
|
|
"language_info": {
|
|
"codemirror_mode": {
|
|
"name": "ipython",
|
|
"version": 3
|
|
},
|
|
"file_extension": ".py",
|
|
"mimetype": "text/x-python",
|
|
"name": "python",
|
|
"nbconvert_exporter": "python",
|
|
"pygments_lexer": "ipython3",
|
|
"version": "3.10.4"
|
|
},
|
|
"orig_nbformat": 4,
|
|
"vscode": {
|
|
"interpreter": {
|
|
"hash": "fddffe4ac3b9f00470127629076101c1b5f38ecb1e7358b567d19305425e9491"
|
|
}
|
|
}
|
|
},
|
|
"nbformat": 4,
|
|
"nbformat_minor": 2
|
|
}
|