Instructions to use jihwan1205/svp-v-coconut-gpt2 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use jihwan1205/svp-v-coconut-gpt2 with Transformers:

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

pipe = pipeline("text-generation", model="jihwan1205/svp-v-coconut-gpt2")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("jihwan1205/svp-v-coconut-gpt2")
model = AutoModelForCausalLM.from_pretrained("jihwan1205/svp-v-coconut-gpt2", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use jihwan1205/svp-v-coconut-gpt2 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "jihwan1205/svp-v-coconut-gpt2"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "jihwan1205/svp-v-coconut-gpt2",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/jihwan1205/svp-v-coconut-gpt2

SGLang

How to use jihwan1205/svp-v-coconut-gpt2 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 "jihwan1205/svp-v-coconut-gpt2" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "jihwan1205/svp-v-coconut-gpt2",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

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 "jihwan1205/svp-v-coconut-gpt2" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "jihwan1205/svp-v-coconut-gpt2",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use jihwan1205/svp-v-coconut-gpt2 with Docker Model Runner:
```
docker model run hf.co/jihwan1205/svp-v-coconut-gpt2
```

jihwan1205 commited on 5 days ago

Commit

c94e146

verified ·

1 Parent(s): 1aa5f3b

Update README.md

Browse files

Files changed (1) hide show

README.md +3 -17

README.md CHANGED Viewed

@@ -28,16 +28,11 @@ Only the value-projection weights (the V columns of each `c_attn`) differ from t
 checkpoint. The perturbation is a training-time exploration mechanism and is **not** used
 at deployment: inference is ordinary greedy decoding.
-This repository contains the paper's GPT-2 SVP-V-GRPO checkpoint from epoch 9
-`B=32`, `lr=6e-5`, seed-0 run. The results below are the paper's Figure 3
-evaluations; they have not been independently rerun for this model-card update.
 ## Results
-Clean greedy decoding, `max_new_tokens=64`, exact-match on the final answer; all rows
-measured in one harness. COCONUT, SLPO, SIM-CoT, and CoDi are the checkpoints released by
-those works; Temp-GRPO / Gaussian-V-GRPO / Dropout-GRPO are GRPO runs from the same
-COCONUT start under the same recipe, differing only in the rollout exploration.
 | Model (GPT-2 124M) | GSM8K | GSM-Hard | MultiArith | SVAMP | ASDiv-A | GSM-Plus |
 |---|---|---|---|---|---|---|
@@ -45,9 +40,6 @@ COCONUT start under the same recipe, differing only in the rollout exploration.
 | SLPO | 34.9 | 7.6 | 82.8 | 34.3 | 58.7 | 18.2 |
 | SIM-CoT | 44.7 | 9.3 | 90.5 | 40.6 | 67.2 | 21.5 |
 | CoDi | 42.5 | 9.3 | 91.9 | 40.0 | 65.4 | 23.1 |
-| Temp-GRPO (answer-token sampling, T=1) | 38.4 | 8.7 | 85.3 | 37.1 | 61.5 | 19.8 |
-| Gaussian-V-GRPO (entry-wise noise on `W_V`) | 40.6 | 9.0 | 88.6 | 39.1 | 62.7 | 21.9 |
-| Dropout-GRPO (native dropout, p=0.2) | 45.2 | 9.3 | 90.9 | **44.6** | 71.1 | 24.8 |
 | **This model** (SVP on `W_V`, α=0.6) | **50.3** | **11.3** | **93.4** | 43.6 | **71.2** | **27.6** |
 ## Usage
@@ -55,8 +47,7 @@ COCONUT start under the same recipe, differing only in the rollout exploration.
 This is a fixed-length continuous-thought model: the prompt ends with
 `<|start-latent|>`, six latent steps feed each step's last hidden state back as the next
 input embedding, `<|end-latent|>` closes the latent phase, and the answer is then decoded
-as ordinary tokens. A plain `generate()` call will **not** reproduce the numbers above —
-the two-phase loop is required.
 ```python
 from transformers import AutoTokenizer, GPT2LMHeadModel
@@ -82,11 +73,6 @@ COCONUT inference code works as well.
 | Optimizer | AdamW, constant `lr=6e-5`, seed 0, no warmup or weight decay, gradient clipping at 1.0; epoch-9 checkpoint from 15 epochs |
 | Reward | Final-answer correctness only |
-## Limitations
-English grade-school arithmetic word problems only; the six benchmarks above are the
-extent of tested generalization.
 ## License
 MIT, inherited from the base checkpoint (which derives from `openai-community/gpt2`).

 checkpoint. The perturbation is a training-time exploration mechanism and is **not** used
 at deployment: inference is ordinary greedy decoding.
+This repository contains the paper's GPT-2 SVP-V-GRPO checkpoint from epoch 9 `B=32`, `lr=6e-5`, seed-0 run.
 ## Results
+Clean greedy decoding, `max_new_tokens=64`, exact-match on the final answer.
 | Model (GPT-2 124M) | GSM8K | GSM-Hard | MultiArith | SVAMP | ASDiv-A | GSM-Plus |
 |---|---|---|---|---|---|---|
 | SLPO | 34.9 | 7.6 | 82.8 | 34.3 | 58.7 | 18.2 |
 | SIM-CoT | 44.7 | 9.3 | 90.5 | 40.6 | 67.2 | 21.5 |
 | CoDi | 42.5 | 9.3 | 91.9 | 40.0 | 65.4 | 23.1 |
 | **This model** (SVP on `W_V`, α=0.6) | **50.3** | **11.3** | **93.4** | 43.6 | **71.2** | **27.6** |
 ## Usage
 This is a fixed-length continuous-thought model: the prompt ends with
 `<|start-latent|>`, six latent steps feed each step's last hidden state back as the next
 input embedding, `<|end-latent|>` closes the latent phase, and the answer is then decoded
+as ordinary tokens.
 ```python
 from transformers import AutoTokenizer, GPT2LMHeadModel
 | Optimizer | AdamW, constant `lr=6e-5`, seed 0, no warmup or weight decay, gradient clipping at 1.0; epoch-9 checkpoint from 15 epochs |
 | Reward | Final-answer correctness only |
 ## License
 MIT, inherited from the base checkpoint (which derives from `openai-community/gpt2`).