|
|
@ -80,7 +80,7 @@
|
|
|
|
"source": [
|
|
|
|
"source": [
|
|
|
|
"MODEL_NAME = \"bigscience/bloom-petals\" # select model you like\n",
|
|
|
|
"MODEL_NAME = \"bigscience/bloom-petals\" # select model you like\n",
|
|
|
|
"NUM_PREFIX_TOKENS = 16\n",
|
|
|
|
"NUM_PREFIX_TOKENS = 16\n",
|
|
|
|
"DEVICE = 'cpu'\n",
|
|
|
|
"DEVICE = 'cuda'\n",
|
|
|
|
"BATCH_SIZE = 16\n",
|
|
|
|
"BATCH_SIZE = 16\n",
|
|
|
|
"LR = 1e-2\n",
|
|
|
|
"LR = 1e-2\n",
|
|
|
|
"WEIGHT_DECAY = 0.0\n",
|
|
|
|
"WEIGHT_DECAY = 0.0\n",
|
|
|
@ -279,11 +279,12 @@
|
|
|
|
]
|
|
|
|
]
|
|
|
|
},
|
|
|
|
},
|
|
|
|
{
|
|
|
|
{
|
|
|
|
|
|
|
|
"attachments": {},
|
|
|
|
"cell_type": "markdown",
|
|
|
|
"cell_type": "markdown",
|
|
|
|
"id": "1bbf014f",
|
|
|
|
"id": "1bbf014f",
|
|
|
|
"metadata": {},
|
|
|
|
"metadata": {},
|
|
|
|
"source": [
|
|
|
|
"source": [
|
|
|
|
"## Beyond soft-propmt tuning\n",
|
|
|
|
"## Beyond soft-prompt tuning\n",
|
|
|
|
"\n",
|
|
|
|
"\n",
|
|
|
|
"Let's try to tune model using adapters in the middle of the model."
|
|
|
|
"Let's try to tune model using adapters in the middle of the model."
|
|
|
|
]
|
|
|
|
]
|
|
|
@ -461,7 +462,7 @@
|
|
|
|
"name": "python",
|
|
|
|
"name": "python",
|
|
|
|
"nbconvert_exporter": "python",
|
|
|
|
"nbconvert_exporter": "python",
|
|
|
|
"pygments_lexer": "ipython3",
|
|
|
|
"pygments_lexer": "ipython3",
|
|
|
|
"version": "3.8.9"
|
|
|
|
"version": "3.8.9 (default, Apr 13 2022, 08:48:07) \n[Clang 13.1.6 (clang-1316.0.21.2.5)]"
|
|
|
|
},
|
|
|
|
},
|
|
|
|
"vscode": {
|
|
|
|
"vscode": {
|
|
|
|
"interpreter": {
|
|
|
|
"interpreter": {
|
|
|
|