Upload folder using huggingface_hub
Browse files- .gitattributes +2 -0
- README.md +61 -0
- SMOLM2Prover-Q4_K_M.gguf +3 -0
- SMOLM2Prover.gguf +3 -0
.gitattributes
CHANGED
|
@@ -33,3 +33,5 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
|
|
| 33 |
*.zip filter=lfs diff=lfs merge=lfs -text
|
| 34 |
*.zst filter=lfs diff=lfs merge=lfs -text
|
| 35 |
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
|
|
|
|
|
|
|
|
| 33 |
*.zip filter=lfs diff=lfs merge=lfs -text
|
| 34 |
*.zst filter=lfs diff=lfs merge=lfs -text
|
| 35 |
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
| 36 |
+
SMOLM2Prover-Q4_K_M.gguf filter=lfs diff=lfs merge=lfs -text
|
| 37 |
+
SMOLM2Prover.gguf filter=lfs diff=lfs merge=lfs -text
|
README.md
ADDED
|
@@ -0,0 +1,61 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# SMOLM2Prover - GGUF Format
|
| 2 |
+
|
| 3 |
+
GGUF quantized version of the SMOLM2Prover model for use with llama.cpp and compatible runtimes.
|
| 4 |
+
|
| 5 |
+
## Model Details
|
| 6 |
+
|
| 7 |
+
- **Original Model**: reaperdoesntknow/SMOLM2Prover
|
| 8 |
+
- **Architecture**: LlamaForCausalLM
|
| 9 |
+
- **Context Length**: 8192 tokens
|
| 10 |
+
- **Embedding Dimension**: 960
|
| 11 |
+
- **Layers**: 32
|
| 12 |
+
- **Head Count**: 15 (Q), 5 (KV) - GQA
|
| 13 |
+
|
| 14 |
+
## Available Files
|
| 15 |
+
|
| 16 |
+
| File | Size | Quantization | Quality |
|
| 17 |
+
|------|------|--------------|---------|
|
| 18 |
+
| `SMOLM2Prover.gguf` | 692M | F16 | Original (no quantization) |
|
| 19 |
+
| `SMOLM2Prover-Q4_K_M.gguf` | 258M | Q4_K_M | Recommended (good quality/size balance) |
|
| 20 |
+
|
| 21 |
+
## Usage
|
| 22 |
+
|
| 23 |
+
### With llama.cpp
|
| 24 |
+
|
| 25 |
+
```bash
|
| 26 |
+
# Run with the quantized model
|
| 27 |
+
./llama-cli -m SMOLM2Prover-Q4_K_M.gguf -p "Your prompt here" -n 256
|
| 28 |
+
```
|
| 29 |
+
|
| 30 |
+
### With Ollama
|
| 31 |
+
|
| 32 |
+
Create a `Modelfile`:
|
| 33 |
+
```
|
| 34 |
+
FROM ./SMOLM2Prover-Q4_K_M.gguf
|
| 35 |
+
```
|
| 36 |
+
|
| 37 |
+
Then:
|
| 38 |
+
```bash
|
| 39 |
+
ollama create smolm2prover -f Modelfile
|
| 40 |
+
ollama run smolm2prover
|
| 41 |
+
```
|
| 42 |
+
|
| 43 |
+
### With LM Studio
|
| 44 |
+
|
| 45 |
+
1. Download `SMOLM2Prover-Q4_K_M.gguf`
|
| 46 |
+
2. Place in LM Studio models folder
|
| 47 |
+
3. Load and chat!
|
| 48 |
+
|
| 49 |
+
## Quantization Details
|
| 50 |
+
|
| 51 |
+
The Q4_K_M quantization uses:
|
| 52 |
+
- Q4_K for most weights
|
| 53 |
+
- Q5_0 fallback for tensors not divisible by 256
|
| 54 |
+
- Q6_K/Q8_0 for some critical layers
|
| 55 |
+
|
| 56 |
+
**Size reduction**: 692M → 258M (63% smaller)
|
| 57 |
+
**BPW**: 5.94 bits per weight
|
| 58 |
+
|
| 59 |
+
## License
|
| 60 |
+
|
| 61 |
+
Same as the original model.
|
SMOLM2Prover-Q4_K_M.gguf
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:55f1d76afef32d2a1c3e1d67cd8f7f464286f60ff146270c9eeb875f95f96bbc
|
| 3 |
+
size 270591136
|
SMOLM2Prover.gguf
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d350a1f1c1ce510879cb06d28850d43afe723dfcd5a3c8113a9de699c4f98ae1
|
| 3 |
+
size 725554336
|