Cleanup: drop SHA256 column + remove stale Files section
Browse files
README.md
CHANGED
|
@@ -32,13 +32,13 @@ as `<protein_coding_region>` instead of DNA).
|
|
| 32 |
|
| 33 |
## Available files
|
| 34 |
|
| 35 |
-
| Quant | File | Size |
|
| 36 |
-
| --- | --- | --- |
|
| 37 |
-
| BF16 | `carbon-500m-bf16.gguf` | 982.2 MB |
|
| 38 |
-
| Q8_0 | `carbon-500m-q8_0.gguf` | 524.6 MB |
|
| 39 |
-
| Q6_K | `carbon-500m-q6_k.gguf` | 443.2 MB |
|
| 40 |
-
| Q5_K_M | `carbon-500m-q5_k_m.gguf` | 405.1 MB |
|
| 41 |
-
| Q4_K_M | `carbon-500m-q4_k_m.gguf` | 369.2 MB |
|
| 42 |
|
| 43 |
K-quants are produced with `--token-embedding-type Q8_0 --output-tensor-type Q8_0`
|
| 44 |
so the sensitive embedding/output tensors stay at Q8_0 precision (important for
|
|
@@ -56,11 +56,6 @@ git clone https://github.com/ggml-org/llama.cpp
|
|
| 56 |
cd llama.cpp && cmake -B build && cmake --build build -j
|
| 57 |
```
|
| 58 |
|
| 59 |
-
## Files
|
| 60 |
-
|
| 61 |
-
| File | Quant | Size |
|
| 62 |
-
|---|---|---|
|
| 63 |
-
| `carbon-500m-bf16.gguf` | bf16 (lossless from source) | 983 MB |
|
| 64 |
|
| 65 |
## Usage
|
| 66 |
|
|
|
|
| 32 |
|
| 33 |
## Available files
|
| 34 |
|
| 35 |
+
| Quant | File | Size |
|
| 36 |
+
| --- | --- | --- |
|
| 37 |
+
| BF16 | `carbon-500m-bf16.gguf` | 982.2 MB |
|
| 38 |
+
| Q8_0 | `carbon-500m-q8_0.gguf` | 524.6 MB |
|
| 39 |
+
| Q6_K | `carbon-500m-q6_k.gguf` | 443.2 MB |
|
| 40 |
+
| Q5_K_M | `carbon-500m-q5_k_m.gguf` | 405.1 MB |
|
| 41 |
+
| Q4_K_M | `carbon-500m-q4_k_m.gguf` | 369.2 MB |
|
| 42 |
|
| 43 |
K-quants are produced with `--token-embedding-type Q8_0 --output-tensor-type Q8_0`
|
| 44 |
so the sensitive embedding/output tensors stay at Q8_0 precision (important for
|
|
|
|
| 56 |
cd llama.cpp && cmake -B build && cmake --build build -j
|
| 57 |
```
|
| 58 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 59 |
|
| 60 |
## Usage
|
| 61 |
|