Instructions to use RL-Forgetting-Experiments-3/qwen2.5-3b-code-sft-replay-uniform-lam1-step102 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use RL-Forgetting-Experiments-3/qwen2.5-3b-code-sft-replay-uniform-lam1-step102 with Transformers:

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

pipe = pipeline("text-generation", model="RL-Forgetting-Experiments-3/qwen2.5-3b-code-sft-replay-uniform-lam1-step102")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("RL-Forgetting-Experiments-3/qwen2.5-3b-code-sft-replay-uniform-lam1-step102")
model = AutoModelForCausalLM.from_pretrained("RL-Forgetting-Experiments-3/qwen2.5-3b-code-sft-replay-uniform-lam1-step102", device_map="auto")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=40)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use RL-Forgetting-Experiments-3/qwen2.5-3b-code-sft-replay-uniform-lam1-step102 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "RL-Forgetting-Experiments-3/qwen2.5-3b-code-sft-replay-uniform-lam1-step102"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "RL-Forgetting-Experiments-3/qwen2.5-3b-code-sft-replay-uniform-lam1-step102",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/RL-Forgetting-Experiments-3/qwen2.5-3b-code-sft-replay-uniform-lam1-step102

SGLang

How to use RL-Forgetting-Experiments-3/qwen2.5-3b-code-sft-replay-uniform-lam1-step102 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 "RL-Forgetting-Experiments-3/qwen2.5-3b-code-sft-replay-uniform-lam1-step102" \
    --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": "RL-Forgetting-Experiments-3/qwen2.5-3b-code-sft-replay-uniform-lam1-step102",
		"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 "RL-Forgetting-Experiments-3/qwen2.5-3b-code-sft-replay-uniform-lam1-step102" \
        --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": "RL-Forgetting-Experiments-3/qwen2.5-3b-code-sft-replay-uniform-lam1-step102",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use RL-Forgetting-Experiments-3/qwen2.5-3b-code-sft-replay-uniform-lam1-step102 with Docker Model Runner:
```
docker model run hf.co/RL-Forgetting-Experiments-3/qwen2.5-3b-code-sft-replay-uniform-lam1-step102
```

smilani commited on 2 days ago

Commit

b12e53e

verified ·

1 Parent(s): 8504704

Complete verified q25_sft_rp_uniform_lam1 final-model delivery

Browse files

Files changed (1) hide show

delivery_manifest.json +107 -0

delivery_manifest.json ADDED Viewed

	@@ -0,0 +1,107 @@

+{
+  "base_model": "Qwen/Qwen2.5-3B",
+  "dataset": "qwen3b_code_sft_data_s300",
+  "dataset_sha256": "22510155770eeff3f05ce0a5d1286f6f051b6baf3480a1276d7dc9da8230d345",
+  "family": "qwen2.5-3b",
+  "files": [
+    {
+      "bytes": 605,
+      "path": "added_tokens.json",
+      "sha256": "58b54bbe36fc752f79a24a271ef66a0a0830054b4dfad94bde757d851968060b",
+      "source": "/mnt/mystorageoutput/projects/rl-forgetting/amlt-results/7209847050.08366-c63126c5-4795/qwen3b_code_ordered_lr1e-5_bs256_seed42_s300_rpuniform_lam1.0_gcr_20260922/final/added_tokens.json"
+    },
+    {
+      "bytes": 2427,
+      "path": "chat_template.jinja",
+      "sha256": "44d5f08f3f72b837eaad09f13a54c1f9f4eb58d75240334548b7fd52a5437fa5",
+      "source": "/mnt/mystorageoutput/projects/rl-forgetting/amlt-results/7209847050.08366-c63126c5-4795/qwen3b_code_ordered_lr1e-5_bs256_seed42_s300_rpuniform_lam1.0_gcr_20260922/final/chat_template.jinja"
+    },
+    {
+      "bytes": 1514,
+      "path": "config.json",
+      "sha256": "33ec9cc401d7ecf57f073a7e25df7c131f37cac38d391e58320509e3f0435e61",
+      "source": "/mnt/mystorageoutput/projects/rl-forgetting/amlt-results/7209847050.08366-c63126c5-4795/qwen3b_code_ordered_lr1e-5_bs256_seed42_s300_rpuniform_lam1.0_gcr_20260922/final/config.json"
+    },
+    {
+      "bytes": 117,
+      "path": "generation_config.json",
+      "sha256": "572b47acfde467f61dc18438b2faf85d1db1ad116c63d93afc23a30634ace770",
+      "source": "/mnt/mystorageoutput/projects/rl-forgetting/amlt-results/7209847050.08366-c63126c5-4795/qwen3b_code_ordered_lr1e-5_bs256_seed42_s300_rpuniform_lam1.0_gcr_20260922/final/generation_config.json"
+    },
+    {
+      "bytes": 1671853,
+      "path": "merges.txt",
+      "sha256": "8831e4f1a044471340f7c0a83d7bd71306a5b867e95fd870f74d0c5308a904d5",
+      "source": "/mnt/mystorageoutput/projects/rl-forgetting/amlt-results/7209847050.08366-c63126c5-4795/qwen3b_code_ordered_lr1e-5_bs256_seed42_s300_rpuniform_lam1.0_gcr_20260922/final/merges.txt"
+    },
+    {
+      "bytes": 4982131536,
+      "path": "model-00001-of-00003.safetensors",
+      "sha256": "5b543586cc17686bd1528c05c74c5a9dbe39c07d9e3921139a46777226f36bd5",
+      "source": "/mnt/mystorageoutput/projects/rl-forgetting/amlt-results/7209847050.08366-c63126c5-4795/qwen3b_code_ordered_lr1e-5_bs256_seed42_s300_rpuniform_lam1.0_gcr_20260922/final/model-00001-of-00003.safetensors"
+    },
+    {
+      "bytes": 4932949336,
+      "path": "model-00002-of-00003.safetensors",
+      "sha256": "b9125802a65ba75b4e1cb171e32f459eefab8d63ae7341998b498066452bf7c7",
+      "source": "/mnt/mystorageoutput/projects/rl-forgetting/amlt-results/7209847050.08366-c63126c5-4795/qwen3b_code_ordered_lr1e-5_bs256_seed42_s300_rpuniform_lam1.0_gcr_20260922/final/model-00002-of-00003.safetensors"
+    },
+    {
+      "bytes": 2428723160,
+      "path": "model-00003-of-00003.safetensors",
+      "sha256": "495882e274455036783d126fe144663986199b741748f5b15ac62926e5d49fca",
+      "source": "/mnt/mystorageoutput/projects/rl-forgetting/amlt-results/7209847050.08366-c63126c5-4795/qwen3b_code_ordered_lr1e-5_bs256_seed42_s300_rpuniform_lam1.0_gcr_20260922/final/model-00003-of-00003.safetensors"
+    },
+    {
+      "bytes": 35618,
+      "path": "model.safetensors.index.json",
+      "sha256": "88a50027f81c9f8a1f2c773b695ccc0a9a47e19107d6c70ccf531f322b5304e6",
+      "source": "/mnt/mystorageoutput/projects/rl-forgetting/amlt-results/7209847050.08366-c63126c5-4795/qwen3b_code_ordered_lr1e-5_bs256_seed42_s300_rpuniform_lam1.0_gcr_20260922/final/model.safetensors.index.json"
+    },
+    {
+      "bytes": 616,
+      "path": "special_tokens_map.json",
+      "sha256": "6676f091c8bc4d1b50146427cfde92073402866b87b6e39223227931b70083e9",
+      "source": "/mnt/mystorageoutput/projects/rl-forgetting/amlt-results/7209847050.08366-c63126c5-4795/qwen3b_code_ordered_lr1e-5_bs256_seed42_s300_rpuniform_lam1.0_gcr_20260922/final/special_tokens_map.json"
+    },
+    {
+      "bytes": 11421896,
+      "path": "tokenizer.json",
+      "sha256": "9c5ae00e602b8860cbd784ba82a8aa14e8feecec692e7076590d014d7b7fdafa",
+      "source": "/mnt/mystorageoutput/projects/rl-forgetting/amlt-results/7209847050.08366-c63126c5-4795/qwen3b_code_ordered_lr1e-5_bs256_seed42_s300_rpuniform_lam1.0_gcr_20260922/final/tokenizer.json"
+    },
+    {
+      "bytes": 4689,
+      "path": "tokenizer_config.json",
+      "sha256": "60f6e7bc948cedd377a20f01d022ef664d026fbc93d72e0a9bdac233f8632181",
+      "source": "/mnt/mystorageoutput/projects/rl-forgetting/amlt-results/7209847050.08366-c63126c5-4795/qwen3b_code_ordered_lr1e-5_bs256_seed42_s300_rpuniform_lam1.0_gcr_20260922/final/tokenizer_config.json"
+    },
+    {
+      "bytes": 2776833,
+      "path": "vocab.json",
+      "sha256": "ca10d7e9fb3ed18575dd1e277a2579c16d108e32f27439684afa0e10b1440910",
+      "source": "/mnt/mystorageoutput/projects/rl-forgetting/amlt-results/7209847050.08366-c63126c5-4795/qwen3b_code_ordered_lr1e-5_bs256_seed42_s300_rpuniform_lam1.0_gcr_20260922/final/vocab.json"
+    },
+    {
+      "bytes": 432,
+      "path": "README.md",
+      "sha256": "05896779dc7a6c56f62767576fc5bd25d4f105c15cd7425f0acd1adc5f4be628",
+      "source": "generated model card"
+    }
+  ],
+  "inference_ready": true,
+  "logical_run": "q25_sft_rp_uniform_lam1",
+  "model_type": "qwen2",
+  "order": "ordered",
+  "replay_lambda": 1.0,
+  "replay_strategy": "uniform",
+  "repo_id": "RL-Forgetting-Experiments-3/qwen2.5-3b-code-sft-replay-uniform-lam1-step102",
+  "schema_version": 1,
+  "source_model": "/mnt/mystorageoutput/projects/rl-forgetting/amlt-results/7209847050.08366-c63126c5-4795/qwen3b_code_ordered_lr1e-5_bs256_seed42_s300_rpuniform_lam1.0_gcr_20260922/final",
+  "source_result_id": "7209847050.08366-c63126c5-4795",
+  "source_run_dir": "qwen3b_code_ordered_lr1e-5_bs256_seed42_s300_rpuniform_lam1.0_gcr_20260922",
+  "status": "complete",
+  "step": 102,
+  "training_state_included": false,
+  "weight_bytes": 12343804032
+}