Skip to content

Run notebook end-to-end with GTX 3090/4090 #5

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
wants to merge 1 commit into
base: main
Choose a base branch
from
Open

Run notebook end-to-end with GTX 3090/4090 #5

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
46 changes: 45 additions & 1 deletion lm-hackers.ipynb
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1256,6 +1256,33 @@
"tokr.batch_decode(res)"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "6e769bff",
"metadata": {},
"outputs": [],
"source": [
"# Funciton to clean memory before using new model\n",
"import gc\n",
"\n",
"# Conveninence function\n",
"def get_cuda_memory_reserved_gb():\n",
" return torch.cuda.memory_reserved()/1_000_000_000\n",
" \n",
"def free_memory(verbose=False):\n",
" gc.collect() \n",
" torch.cuda.empty_cache()\n",
" if verbose:\n",
" print(f'ℹ️ CUDA MEMORY RESERVED AFTER free_memory: {get_cuda_memory_reserved_gb()} GB') \n",
"\n",
"# Let's see how it works:\n",
"print(f'BREFORE free_memory: {get_cuda_memory_reserved_gb()} GB')\n",
"del model # Delete model to be able to free memory\n",
"print(f'BREFORE free_memory after deleting `model`: {get_cuda_memory_reserved_gb()} GB')\n",
"free_memory(verbose=True)"
]
},
{
"cell_type": "code",
"execution_count": 28,
Expand Down Expand Up @@ -1321,7 +1348,9 @@
"metadata": {},
"outputs": [],
"source": [
"model = AutoModelForCausalLM.from_pretrained('TheBloke/Llama-2-7b-Chat-GPTQ', device_map=0, torch_dtype=torch.float16)"
"del model; free_memory(True); # Cleanup before loading new one\n",
"model = AutoModelForCausalLM.from_pretrained('TheBloke/Llama-2-7b-Chat-GPTQ', device_map=0, torch_dtype=torch.float16)\n",
"print(f'CUDA MEMORY RESERVED AFTER LOADING model: {get_cuda_memory_reserved_gb()} GB')"
]
},
{
Expand Down Expand Up @@ -1364,6 +1393,7 @@
"metadata": {},
"outputs": [],
"source": [
"del model; free_memory();\n",
"mn = 'TheBloke/Llama-2-13B-GPTQ'\n",
"model = AutoModelForCausalLM.from_pretrained(mn, device_map=0, torch_dtype=torch.float16)"
]
Expand Down Expand Up @@ -1465,6 +1495,7 @@
}
],
"source": [
"del model; free_memory();\n",
"mn = \"stabilityai/StableBeluga-7B\"\n",
"model = AutoModelForCausalLM.from_pretrained(mn, device_map=0, torch_dtype=torch.bfloat16)"
]
Expand Down Expand Up @@ -1535,6 +1566,7 @@
"metadata": {},
"outputs": [],
"source": [
"del model; free_memory();\n",
"mn = 'TheBloke/OpenOrca-Platypus2-13B-GPTQ'\n",
"model = AutoModelForCausalLM.from_pretrained(mn, device_map=0, torch_dtype=torch.float16)"
]
Expand Down Expand Up @@ -2069,6 +2101,7 @@
"metadata": {},
"outputs": [],
"source": [
"del model; free_memory();\n",
"model = AutoModelForCausalLM.from_pretrained('meta-llama/Llama-2-7b-hf',\n",
" torch_dtype=torch.bfloat16, device_map=0)\n",
"model = PeftModel.from_pretrained(model, ax_model)\n",
Expand Down Expand Up @@ -2119,6 +2152,17 @@
"print(tokr.batch_decode(res)[0])"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "b17c43c0",
"metadata": {},
"outputs": [],
"source": [
"# Finally release model and memory\n",
"del model; free_memory(True);"
]
},
{
"cell_type": "markdown",
"id": "991d4a93-dab8-4777-82d2-301c494deba0",
Expand Down