Instructions to use LayerEight/Cumulus-Qwen2.5-7B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use LayerEight/Cumulus-Qwen2.5-7B with PEFT:

from peft import PeftModel
from transformers import AutoModelForCausalLM

base_model = AutoModelForCausalLM.from_pretrained("Qwen/Qwen2.5-7B-Instruct")
model = PeftModel.from_pretrained(base_model, "LayerEight/Cumulus-Qwen2.5-7B")

Transformers

How to use LayerEight/Cumulus-Qwen2.5-7B with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="LayerEight/Cumulus-Qwen2.5-7B")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoModel
model = AutoModel.from_pretrained("LayerEight/Cumulus-Qwen2.5-7B", dtype="auto")

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use LayerEight/Cumulus-Qwen2.5-7B with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "LayerEight/Cumulus-Qwen2.5-7B"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "LayerEight/Cumulus-Qwen2.5-7B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/LayerEight/Cumulus-Qwen2.5-7B

SGLang

How to use LayerEight/Cumulus-Qwen2.5-7B with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "LayerEight/Cumulus-Qwen2.5-7B" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "LayerEight/Cumulus-Qwen2.5-7B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "LayerEight/Cumulus-Qwen2.5-7B" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "LayerEight/Cumulus-Qwen2.5-7B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use LayerEight/Cumulus-Qwen2.5-7B with Docker Model Runner:
```
docker model run hf.co/LayerEight/Cumulus-Qwen2.5-7B
```

LayerEight commited on 13 days ago

Commit

97c89fd

verified ·

1 Parent(s): d4b08fe

Upload folder using huggingface_hub

Browse files

Files changed (6) hide show

adapter_config.json +2 -2
adapter_model.safetensors +1 -1
checkpoint-77/adapter_config.json +2 -2
checkpoint-77/adapter_model.safetensors +1 -1
checkpoint-77/optimizer.pt +1 -1
checkpoint-77/trainer_state.json +12 -12

adapter_config.json CHANGED Viewed

@@ -30,12 +30,12 @@
   "rank_pattern": {},
   "revision": null,
   "target_modules": [
-    "v_proj",
     "o_proj",
     "up_proj",
     "q_proj",
     "gate_proj",
-    "down_proj",
     "k_proj"
   ],
   "target_parameters": null,

   "rank_pattern": {},
   "revision": null,
   "target_modules": [
     "o_proj",
     "up_proj",
+    "v_proj",
+    "down_proj",
     "q_proj",
     "gate_proj",
     "k_proj"
   ],
   "target_parameters": null,

adapter_model.safetensors CHANGED Viewed

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:4c30d61139d1d4f602eaa8cc55ffa095bda15571a7ca8410511e9c1a13818115
 size 80792880

 version https://git-lfs.github.com/spec/v1
+oid sha256:828f20aa28d6fde2f54ec3cef68c52c86b74ba89af6430c280bfddcf66de93d0
 size 80792880

checkpoint-77/adapter_config.json CHANGED Viewed

@@ -30,12 +30,12 @@
   "rank_pattern": {},
   "revision": null,
   "target_modules": [
-    "v_proj",
     "o_proj",
     "up_proj",
     "q_proj",
     "gate_proj",
-    "down_proj",
     "k_proj"
   ],
   "target_parameters": null,

   "rank_pattern": {},
   "revision": null,
   "target_modules": [
     "o_proj",
     "up_proj",
+    "v_proj",
+    "down_proj",
     "q_proj",
     "gate_proj",
     "k_proj"
   ],
   "target_parameters": null,

checkpoint-77/adapter_model.safetensors CHANGED Viewed

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:4c30d61139d1d4f602eaa8cc55ffa095bda15571a7ca8410511e9c1a13818115
 size 80792880

 version https://git-lfs.github.com/spec/v1
+oid sha256:828f20aa28d6fde2f54ec3cef68c52c86b74ba89af6430c280bfddcf66de93d0
 size 80792880

checkpoint-77/optimizer.pt CHANGED Viewed

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a8f1f2b1b61f776e2b88b9c5fe6a777fcdc76352d7526f95d95fe4328be603d0
 size 161810747

 version https://git-lfs.github.com/spec/v1
+oid sha256:2cb711327043f058babaf737266d6285bc02546ee7853279764ac08b99c80b26
 size 161810747

checkpoint-77/trainer_state.json CHANGED Viewed

@@ -10,32 +10,32 @@
   "is_world_process_zero": true,
   "log_history": [
     {
-      "entropy": 1.2584150770306588,
       "epoch": 0.32786885245901637,
-      "grad_norm": 0.053955078125,
       "learning_rate": 0.00016282199972956425,
-      "loss": 1.263802261352539,
-      "mean_token_accuracy": 0.7028659734129906,
       "num_tokens": 150039.0,
       "step": 25
     },
     {
-      "entropy": 0.9214049679040909,
       "epoch": 0.6557377049180327,
-      "grad_norm": 0.1962890625,
       "learning_rate": 6.271435222196916e-05,
-      "loss": 0.829261474609375,
-      "mean_token_accuracy": 0.7762054482102394,
       "num_tokens": 292991.0,
       "step": 50
     },
     {
-      "entropy": 0.620386914536357,
       "epoch": 0.9836065573770492,
-      "grad_norm": 0.1064453125,
       "learning_rate": 8.099564741123166e-07,
-      "loss": 0.5387606430053711,
-      "mean_token_accuracy": 0.8512335225939751,
       "num_tokens": 440288.0,
       "step": 75
     }

   "is_world_process_zero": true,
   "log_history": [
     {
+      "entropy": 1.2594475726783276,
       "epoch": 0.32786885245901637,
+      "grad_norm": 0.05419921875,
       "learning_rate": 0.00016282199972956425,
+      "loss": 1.2619928741455078,
+      "mean_token_accuracy": 0.7029473577439785,
       "num_tokens": 150039.0,
       "step": 25
     },
     {
+      "entropy": 0.923663822337985,
       "epoch": 0.6557377049180327,
+      "grad_norm": 0.19140625,
       "learning_rate": 6.271435222196916e-05,
+      "loss": 0.82879150390625,
+      "mean_token_accuracy": 0.776209542453289,
       "num_tokens": 292991.0,
       "step": 50
     },
     {
+      "entropy": 0.6226911177486181,
       "epoch": 0.9836065573770492,
+      "grad_norm": 0.103515625,
       "learning_rate": 8.099564741123166e-07,
+      "loss": 0.5372924423217773,
+      "mean_token_accuracy": 0.8512885943055153,
       "num_tokens": 440288.0,
       "step": 75
     }