diff --git a/.github/workflows/stale.yml b/.github/workflows/stale.yml
new file mode 100644
index 0000000..22706e8
--- /dev/null
+++ b/.github/workflows/stale.yml
@@ -0,0 +1,30 @@
+name: "Mark and close stale issues"
+on:
+ workflow_dispatch:
+ schedule:
+ - cron: "0 0 * * *"
+
+jobs:
+ stale:
+ if: ${{ github.repository == 'deepseek-ai/DeepSeek-V3' }}
+ runs-on: ubuntu-latest
+ steps:
+ - name: "Mark and close stale issues"
+ uses: actions/stale@v9
+ with:
+ days-before-issue-stale: 30
+ days-before-issue-close: 14
+ stale-issue-label: "stale"
+ close-issue-label: "closed-as-stale"
+ exempt-issue-labels: |
+ pinned
+ security
+ stale-issue-message: >
+ This issue has been automatically marked as stale because it has not had
+ recent activity. It will be closed if no further activity occurs. If you
+ believe this issue is still relevant, please leave a comment to keep it open.
+ Thank you for your contributions!
+ close-issue-message: false
+ days-before-pr-stale: -1
+ days-before-pr-close: -1
+ repo-token: ${{ secrets.GITHUB_TOKEN }}
diff --git a/README.md b/README.md
index 7ecf87e..9ba2346 100644
--- a/README.md
+++ b/README.md
@@ -7,42 +7,39 @@
-
-
-
-
-
-
-
+ 
+ 
+ 
+
+ 
+ 
+ 
+
+ 
+ 
+
Paper Link👁️
-
+
+
+## Table of Contents
+
+1. [Introduction](#1-introduction)
+2. [Model Summary](#2-model-summary)
+3. [Model Downloads](#3-model-downloads)
+4. [Evaluation Results](#4-evaluation-results)
+5. [Chat Website & API Platform](#5-chat-website--api-platform)
+6. [How to Run Locally](#6-how-to-run-locally)
+7. [License](#7-license)
+8. [Citation](#8-citation)
+9. [Contact](#9-contact)
## 1. Introduction
@@ -99,7 +96,7 @@ Throughout the entire training process, we did not experience any irrecoverable
> [!NOTE]
-> The total size of DeepSeek-V3 models on Hugging Face is 685B, which includes 671B of the Main Model weights and 14B of the Multi-Token Prediction (MTP) Module weights.**
+> The total size of DeepSeek-V3 models on Hugging Face is 685B, which includes 671B of the Main Model weights and 14B of the Multi-Token Prediction (MTP) Module weights.
To ensure optimal performance and flexibility, we have partnered with open-source communities and hardware vendors to provide multiple ways to run the model locally. For step-by-step guidance, check out Section 6: [How_to Run_Locally](#6-how-to-run-locally).
@@ -130,7 +127,7 @@ For developers looking to dive deeper, we recommend exploring [README_WEIGHTS.md
| | WinoGrande (Acc.) | 5-shot | **86.3** | 82.3 | 85.2 | 84.9 |
| | RACE-Middle (Acc.) | 5-shot | 73.1 | 68.1 | **74.2** | 67.1 |
| | RACE-High (Acc.) | 5-shot | 52.6 | 50.3 | **56.8** | 51.3 |
-| | TriviaQA (EM) | 5-shot | 80.0 | 71.9 | **82.7** | **82.9** |
+| | TriviaQA (EM) | 5-shot | 80.0 | 71.9 | 82.7 | **82.9** |
| | NaturalQuestions (EM) | 5-shot | 38.6 | 33.2 | **41.5** | 40.0 |
| | AGIEval (Acc.) | 0-shot | 57.5 | 75.8 | 60.6 | **79.6** |
| Code | HumanEval (Pass@1) | 0-shot | 43.3 | 53.0 | 54.9 | **65.2** |
@@ -249,7 +246,7 @@ python fp8_cast_bf16.py --input-fp8-hf-path /path/to/fp8_weights --output-bf16-h
```
> [!NOTE]
-> Hugging Face's Transformers has not been directly supported yet.**
+> Hugging Face's Transformers has not been directly supported yet.
### 6.1 Inference with DeepSeek-Infer Demo (example only)
@@ -259,7 +256,7 @@ python fp8_cast_bf16.py --input-fp8-hf-path /path/to/fp8_weights --output-bf16-h
> Linux with Python 3.10 only. Mac and Windows are not supported.
Dependencies:
-```
+```pip-requirements
torch==2.4.1
triton==3.0.0
transformers==4.46.3
diff --git a/inference/convert.py b/inference/convert.py
index c606ce8..6d85ccc 100644
--- a/inference/convert.py
+++ b/inference/convert.py
@@ -60,7 +60,7 @@ def main(hf_ckpt_path, save_path, n_experts, mp):
name = name.replace("weight_scale_inv", "scale")
name = name.replace("e_score_correction_bias", "bias")
key = name.split(".")[-2]
- assert key in mapping
+ assert key in mapping, f"Key {key} not found in mapping"
new_key, dim = mapping[key]
name = name.replace(key, new_key)
for i in range(mp):
@@ -70,7 +70,7 @@ def main(hf_ckpt_path, save_path, n_experts, mp):
if idx < i * n_local_experts or idx >= (i + 1) * n_local_experts:
continue
elif dim is not None:
- assert param.size(dim) % mp == 0
+ assert param.size(dim) % mp == 0, f"Dimension {dim} must be divisible by {mp}"
shard_size = param.size(dim) // mp
new_param = param.narrow(dim, i * shard_size, shard_size).contiguous()
state_dicts[i][name] = new_param
@@ -92,5 +92,5 @@ if __name__ == "__main__":
parser.add_argument("--n-experts", type=int, required=True)
parser.add_argument("--model-parallel", type=int, required=True)
args = parser.parse_args()
- assert args.n_experts % args.model_parallel == 0
+ assert args.n_experts % args.model_parallel == 0, "Number of experts must be divisible by model parallelism"
main(args.hf_ckpt_path, args.save_path, args.n_experts, args.model_parallel)
diff --git a/inference/generate.py b/inference/generate.py
index fbf3ab8..7e9bffe 100644
--- a/inference/generate.py
+++ b/inference/generate.py
@@ -49,7 +49,7 @@ def generate(
List[List[int]]: A list of lists containing the generated tokens for each sequence.
"""
prompt_lens = [len(t) for t in prompt_tokens]
- assert max(prompt_lens) <= model.max_seq_len
+ assert max(prompt_lens) <= model.max_seq_len, f"Prompt length exceeds model maximum sequence length (max_seq_len={model.max_seq_len})"
total_len = min(model.max_seq_len, max_new_tokens + max(prompt_lens))
tokens = torch.full((len(prompt_tokens), total_len), -1, dtype=torch.long, device="cuda")
for i, t in enumerate(prompt_tokens):
@@ -145,7 +145,7 @@ def main(
else:
with open(input_file) as f:
prompts = [line.strip() for line in f.readlines()]
- assert len(prompts) <= args.max_batch_size
+ assert len(prompts) <= args.max_batch_size, f"Number of prompts exceeds maximum batch size ({args.max_batch_size})"
prompt_tokens = [tokenizer.apply_chat_template([{"role": "user", "content": prompt}], add_generation_prompt=True) for prompt in prompts]
completion_tokens = generate(model, prompt_tokens, max_new_tokens, tokenizer.eos_token_id, temperature)
completions = tokenizer.batch_decode(completion_tokens, skip_special_tokens=True)
@@ -181,5 +181,5 @@ if __name__ == "__main__":
parser.add_argument("--max-new-tokens", type=int, default=200)
parser.add_argument("--temperature", type=float, default=0.2)
args = parser.parse_args()
- assert args.input_file or args.interactive
+ assert args.input_file or args.interactive, "Either input-file or interactive mode must be specified"
main(args.ckpt_path, args.config, args.input_file, args.interactive, args.max_new_tokens, args.temperature)
diff --git a/inference/kernel.py b/inference/kernel.py
index dec8639..ae907ad 100644
--- a/inference/kernel.py
+++ b/inference/kernel.py
@@ -43,8 +43,8 @@ def act_quant(x: torch.Tensor, block_size: int = 128) -> Tuple[torch.Tensor, tor
- The quantized tensor with dtype `torch.float8_e4m3fn`.
- A tensor of scaling factors with dtype `torch.float32`.
"""
- assert x.is_contiguous()
- assert x.size(-1) % block_size == 0
+ assert x.is_contiguous(), 'Input tensor must be contiguous'
+ assert x.size(-1) % block_size == 0, f'Last dimension size must be divisible by block_size (block_size={block_size})'
y = torch.empty_like(x, dtype=torch.float8_e4m3fn)
s = x.new_empty(*x.size()[:-1], x.size(-1) // block_size, dtype=torch.float32)
grid = lambda meta: (triton.cdiv(x.numel(), meta['BLOCK_SIZE']), )
@@ -96,8 +96,8 @@ def weight_dequant(x: torch.Tensor, s: torch.Tensor, block_size: int = 128) -> t
Raises:
AssertionError: If `x` or `s` are not contiguous or if their dimensions are not 2.
"""
- assert x.is_contiguous() and s.is_contiguous()
- assert x.dim() == 2 and s.dim() == 2
+ assert x.is_contiguous() and s.is_contiguous(), 'Input tensors must be contiguous'
+ assert x.dim() == 2 and s.dim() == 2, 'Input tensors must have 2 dimensions'
M, N = x.size()
y = torch.empty_like(x, dtype=torch.get_default_dtype())
grid = lambda meta: (triton.cdiv(M, meta['BLOCK_SIZE']), triton.cdiv(N, meta['BLOCK_SIZE']))
@@ -180,8 +180,8 @@ def fp8_gemm(a: torch.Tensor, a_s: torch.Tensor, b: torch.Tensor, b_s: torch.Ten
Returns:
torch.Tensor: The result of the matrix multiplication.
"""
- assert a.is_contiguous() and b.is_contiguous()
- assert a_s.is_contiguous() and b_s.is_contiguous()
+ assert a.is_contiguous() and b.is_contiguous(), 'Input tensors must be contiguous'
+ assert a_s.is_contiguous() and b_s.is_contiguous(), 'Scaling factor tensors must be contiguous'
K = a.size(-1)
M = a.numel() // K
N = b.size(0)
diff --git a/inference/model.py b/inference/model.py
index 29e2931..ad09129 100644
--- a/inference/model.py
+++ b/inference/model.py
@@ -89,7 +89,7 @@ class ParallelEmbedding(nn.Module):
super().__init__()
self.vocab_size = vocab_size
self.dim = dim
- assert vocab_size % world_size == 0
+ assert vocab_size % world_size == 0, f"Vocabulary size must be divisible by world size (world_size={world_size})"
self.part_vocab_size = (vocab_size // world_size)
self.vocab_start_idx = rank * self.part_vocab_size
self.vocab_end_idx = self.vocab_start_idx + self.part_vocab_size
@@ -124,7 +124,7 @@ def linear(x: torch.Tensor, weight: torch.Tensor, bias: Optional[torch.Tensor] =
quantization-aware computations depending on the input parameters.
Notes:
- - If `weight` is quantized (e.g., `element_size() > 1`), a dequantized version
+ - If `weight` is quantized (e.g., `element_size() == 1`), a dequantized version
is used for computation.
- If `gemm_impl == "bf16"`, dequantization and a `bf16` GEMM operation are applied.
- For other cases, the function applies quantization to `x` and uses `fp8_gemm` for computation.
@@ -176,7 +176,7 @@ class ColumnParallelLinear(Linear):
dtype (optional): Data type for the layer. Defaults to `torch.bfloat16`.
"""
def __init__(self, in_features: int, out_features: int, bias: bool = False, dtype = None):
- assert out_features % world_size == 0
+ assert out_features % world_size == 0, f"Output features must be divisible by world size (world_size={world_size})"
self.part_out_features = out_features // world_size
super().__init__(in_features, self.part_out_features, bias, dtype)
@@ -205,7 +205,7 @@ class RowParallelLinear(Linear):
dtype (optional): Data type for the layer. Defaults to `torch.bfloat16`.
"""
def __init__(self, in_features: int, out_features: int, bias: bool = False, dtype = None):
- assert in_features % world_size == 0
+ assert in_features % world_size == 0, f"Input features must be divisible by world size (world_size={world_size})"
self.part_in_features = in_features // world_size
super().__init__(self.part_in_features, out_features, bias, dtype)
@@ -566,8 +566,8 @@ class Gate(nn.Module):
else:
group_scores = scores.topk(2, dim=-1)[0].sum(dim=-1)
indices = group_scores.topk(self.topk_groups, dim=-1)[1]
- mask = torch.zeros_like(scores[..., 0]).scatter_(1, indices, True)
- scores = (scores * mask.unsqueeze(-1)).flatten(1)
+ mask = scores.new_ones(x.size(0), self.n_groups, dtype=bool).scatter_(1, indices, False)
+ scores = scores.masked_fill_(mask.unsqueeze(-1), float("-inf")).flatten(1)
indices = torch.topk(scores, self.topk, dim=-1)[1]
weights = original_scores.gather(1, indices)
if self.score_func == "sigmoid":
@@ -633,7 +633,7 @@ class MoE(nn.Module):
"""
super().__init__()
self.dim = args.dim
- assert args.n_routed_experts % world_size == 0
+ assert args.n_routed_experts % world_size == 0, f"Number of experts must be divisible by world size (world_size={world_size})"
self.n_routed_experts = args.n_routed_experts
self.n_local_experts = args.n_routed_experts // world_size
self.n_activated_experts = args.n_activated_experts
