Bendi新闻
>
Karpathy称赞，从零实现LLaMa3项目爆火，半天1.5k star
Karpathy称赞，从零实现LLaMa3项目爆火，半天1.5k star
7月前
编辑：杜伟、陈萍
项目中代码很多很全，值得细读。
from pathlib import Pathimport tiktokenfrom tiktoken.load import load_tiktoken_bpeimport torchimport jsonimport matplotlib.pyplot as plttokenizer_path = "Meta-Llama-3-8B/tokenizer.model"special_tokens = ["<|begin_of_text|>","<|end_of_text|>","<|reserved_special_token_0|>","<|reserved_special_token_1|>","<|reserved_special_token_2|>","<|reserved_special_token_3|>","<|start_header_id|>","<|end_header_id|>","<|reserved_special_token_4|>","<|eot_id|>", # end of turn] + [f"<|reserved_special_token_{i}|>" for i in range (5, 256 - 5)] mergeable_ranks = load_tiktoken_bpe (tokenizer_path) tokenizer = tiktoken.Encoding (name=Path (tokenizer_path).name,pat_str=r"(?i:'s|'t|'re|'ve|'m|'ll|'d)|[^\r\n\p {L}\p {N}]?\p {L}+|\p {N}{1,3}| ?[^\s\p {L}\p {N}]+[\r\n]*|\s*[\r\n]+|\s+(?!\S)|\s+",mergeable_ranks=mergeable_ranks,special_tokens={token: len (mergeable_ranks) + i for i, token in enumerate (special_tokens)},)tokenizer.decode (tokenizer.encode ("hello world!"))
'hello world!'model = torch.load ("Meta-Llama-3-8B/consolidated.00.pth")print (json.dumps (list (model.keys ())[:20], indent=4))
["tok_embeddings.weight","layers.0.attention.wq.weight","layers.0.attention.wk.weight","layers.0.attention.wv.weight","layers.0.attention.wo.weight","layers.0.feed_forward.w1.weight","layers.0.feed_forward.w3.weight","layers.0.feed_forward.w2.weight","layers.0.attention_norm.weight","layers.0.ffn_norm.weight","layers.1.attention.wq.weight","layers.1.attention.wk.weight","layers.1.attention.wv.weight","layers.1.attention.wo.weight","layers.1.feed_forward.w1.weight","layers.1.feed_forward.w3.weight","layers.1.feed_forward.w2.weight","layers.1.attention_norm.weight","layers.1.ffn_norm.weight","layers.2.attention.wq.weight"]
with open ("Meta-Llama-3-8B/params.json", "r") as f:config = json.load (f)config
{'dim': 4096,'n_layers': 32,'n_heads': 32,'n_kv_heads': 8,'vocab_size': 128256,'multiple_of': 1024,'ffn_dim_multiplier': 1.3,'norm_eps': 1e-05,'rope_theta': 500000.0}
模型有 32 个 transformer 层； 每个多头注意力块有 32 个头。
dim = config ["dim"]n_layers = config ["n_layers"]n_heads = config ["n_heads"]n_kv_heads = config ["n_kv_heads"]vocab_size = config ["vocab_size"]multiple_of = config ["multiple_of"]ffn_dim_multiplier = config ["ffn_dim_multiplier"]norm_eps = config ["norm_eps"]rope_theta = torch.tensor (config ["rope_theta"])
prompt = "the answer to the ultimate question of life, the universe, and everything is"tokens = [128000] + tokenizer.encode (prompt)print (tokens)tokens = torch.tensor (tokens)prompt_split_as_tokens = [tokenizer.decode ([token.item ()]) for token in tokens]print (prompt_split_as_tokens)
[128000, 1820, 4320, 311, 279, 17139, 3488, 315, 2324, 11, 279, 15861, 11, 323, 4395, 374, 220]['<|begin_of_text|>', 'the', ' answer', ' to', ' the', ' ultimate', ' question', ' of', ' life', ',', ' the', ' universe', ',', ' and', ' everything', ' is', ' ']
embedding_layer = torch.nn.Embedding (vocab_size, dim)embedding_layer.weight.data.copy_(model ["tok_embeddings.weight"])token_embeddings_unnormalized = embedding_layer (tokens).to (torch.bfloat16)token_embeddings_unnormalized.shape
torch.Size ([17, 4096])# def rms_norm (tensor, norm_weights):# rms = (tensor.pow (2).mean (-1, keepdim=True) + norm_eps)**0.5# return tensor * (norm_weights /rms)def rms_norm (tensor, norm_weights):return (tensor * torch.rsqrt (tensor.pow (2).mean (-1, keepdim=True) + norm_eps)) * norm_weights
token_embeddings = rms_norm (token_embeddings_unnormalized, model ["layers.0.attention_norm.weight"])token_embeddings.shape
torch.Size ([17, 4096])print (model ["layers.0.attention.wq.weight"].shape,model ["layers.0.attention.wk.weight"].shape,model ["layers.0.attention.wv.weight"].shape,model ["layers.0.attention.wo.weight"].shape)torch.Size ([4096, 4096]) torch.Size ([1024, 4096]) torch.Size ([1024, 4096]) torch.Size ([4096, 4096])
q_layer0 = model ["layers.0.attention.wq.weight"]head_dim = q_layer0.shape [0] //n_headsq_layer0 = q_layer0.view (n_heads, head_dim, dim)q_layer0.shape
torch.Size ([32, 128, 4096])q_layer0_head0 = q_layer0 [0]q_layer0_head0.shape
torch.Size ([128, 4096])q_per_token = torch.matmul (token_embeddings, q_layer0_head0.T)q_per_token.shape
torch.Size ([17, 128])q_per_token_split_into_pairs = q_per_token.float ().view (q_per_token.shape [0], -1, 2)q_per_token_split_into_pairs.shape
torch.Size ([17, 64, 2])zero_to_one_split_into_64_parts = torch.tensor (range (64))/64zero_to_one_split_into_64_parts
tensor ([0.0000, 0.0156, 0.0312, 0.0469, 0.0625, 0.0781, 0.0938, 0.1094, 0.1250,0.1406, 0.1562, 0.1719, 0.1875, 0.2031, 0.2188, 0.2344, 0.2500, 0.2656,0.2812, 0.2969, 0.3125, 0.3281, 0.3438, 0.3594, 0.3750, 0.3906, 0.4062,0.4219, 0.4375, 0.4531, 0.4688, 0.4844, 0.5000, 0.5156, 0.5312, 0.5469,0.5625, 0.5781, 0.5938, 0.6094, 0.6250, 0.6406, 0.6562, 0.6719, 0.6875,0.7031, 0.7188, 0.7344, 0.7500, 0.7656, 0.7812, 0.7969, 0.8125, 0.8281,0.8438, 0.8594, 0.8750, 0.8906, 0.9062, 0.9219, 0.9375, 0.9531, 0.9688,0.9844])
freqs = 1.0 / (rope_theta ** zero_to_one_split_into_64_parts)freqs
tensor ([1.0000e+00, 8.1462e-01, 6.6360e-01, 5.4058e-01, 4.4037e-01, 3.5873e-01,2.9223e-01, 2.3805e-01, 1.9392e-01, 1.5797e-01, 1.2869e-01, 1.0483e-01,8.5397e-02, 6.9566e-02, 5.6670e-02, 4.6164e-02, 3.7606e-02, 3.0635e-02,2.4955e-02, 2.0329e-02, 1.6560e-02, 1.3490e-02, 1.0990e-02, 8.9523e-03,7.2927e-03, 5.9407e-03, 4.8394e-03, 3.9423e-03, 3.2114e-03, 2.6161e-03,2.1311e-03, 1.7360e-03, 1.4142e-03, 1.1520e-03, 9.3847e-04, 7.6450e-04,6.2277e-04, 5.0732e-04, 4.1327e-04, 3.3666e-04, 2.7425e-04, 2.2341e-04,1.8199e-04, 1.4825e-04, 1.2077e-04, 9.8381e-05, 8.0143e-05, 6.5286e-05,5.3183e-05, 4.3324e-05, 3.5292e-05, 2.8750e-05, 2.3420e-05, 1.9078e-05,1.5542e-05, 1.2660e-05, 1.0313e-05, 8.4015e-06, 6.8440e-06, 5.5752e-06,4.5417e-06, 3.6997e-06, 3.0139e-06, 2.4551e-06])
freqs_for_each_token = torch.outer (torch.arange (17), freqs)freqs_cis = torch.polar (torch.ones_like (freqs_for_each_token), freqs_for_each_token)freqs_cis.shapevalue = freqs_cis [3]plt.figure ()for i, element in enumerate (value [:17]):plt.plot ([0, element.real], [0, element.imag], color='blue', linewidth=1, label=f"Index: {i}")plt.annotate (f"{i}", xy=(element.real, element.imag), color='red')plt.xlabel ('Real')plt.ylabel ('Imaginary')plt.title ('Plot of one row of freqs_cis')plt.show ()
q_per_token_as_complex_numbers = torch.view_as_complex (q_per_token_split_into_pairs)q_per_token_as_complex_numbers.shape
torch.Size ([17, 64])q_per_token_as_complex_numbers_rotated = q_per_token_as_complex_numbers * freqs_cisq_per_token_as_complex_numbers_rotated.shape
torch.Size ([17, 64])q_per_token_split_into_pairs_rotated = torch.view_as_real (q_per_token_as_complex_numbers_rotated)q_per_token_split_into_pairs_rotated.shape
torch.Size ([17, 64, 2])q_per_token_rotated = q_per_token_split_into_pairs_rotated.view (q_per_token.shape)q_per_token_rotated.shape
torch.Size ([17, 128])k_layer0 = model ["layers.0.attention.wk.weight"]k_layer0 = k_layer0.view (n_kv_heads, k_layer0.shape [0] //n_kv_heads, dim)k_layer0.shape
torch.Size ([8, 128, 4096])k_layer0_head0 = k_layer0 [0]k_layer0_head0.shape
torch.Size ([128, 4096])k_per_token = torch.matmul (token_embeddings, k_layer0_head0.T)k_per_token.shape
torch.Size ([17, 128])k_per_token_split_into_pairs = k_per_token.float ().view (k_per_token.shape [0], -1, 2)k_per_token_split_into_pairs.shape
torch.Size ([17, 64, 2])k_per_token_as_complex_numbers = torch.view_as_complex (k_per_token_split_into_pairs)k_per_token_as_complex_numbers.shape
torch.Size ([17, 64])k_per_token_split_into_pairs_rotated = torch.view_as_real (k_per_token_as_complex_numbers * freqs_cis)k_per_token_split_into_pairs_rotated.shape
torch.Size ([17, 64, 2])k_per_token_rotated = k_per_token_split_into_pairs_rotated.view (k_per_token.shape)k_per_token_rotated.shape
torch.Size ([17, 128])qk_per_token = torch.matmul (q_per_token_rotated, k_per_token_rotated.T)/(head_dim)**0.5qk_per_token.shape
torch.Size ([17, 17])def display_qk_heatmap (qk_per_token):_, ax = plt.subplots ()im = ax.imshow (qk_per_token.to (float).detach (), cmap='viridis')ax.set_xticks (range (len (prompt_split_as_tokens)))ax.set_yticks (range (len (prompt_split_as_tokens)))ax.set_xticklabels (prompt_split_as_tokens)ax.set_yticklabels (prompt_split_as_tokens)ax.figure.colorbar (im, ax=ax)display_qk_heatmap (qk_per_token)
mask = torch.full ((len (tokens), len (tokens)), float ("-inf"), device=tokens.device) mask = torch.triu (mask, diagonal=1) masktensor ([[0., -inf, -inf, -inf, -inf, -inf, -inf, -inf, -inf, -inf, -inf, -inf, -inf, -inf, -inf, -inf, -inf],[0., 0., -inf, -inf, -inf, -inf, -inf, -inf, -inf, -inf, -inf, -inf, -inf, -inf, -inf, -inf, -inf],[0., 0., 0., -inf, -inf, -inf, -inf, -inf, -inf, -inf, -inf, -inf, -inf, -inf, -inf, -inf, -inf],[0., 0., 0., 0., -inf, -inf, -inf, -inf, -inf, -inf, -inf, -inf, -inf, -inf, -inf, -inf, -inf],[0., 0., 0., 0., 0., -inf, -inf, -inf, -inf, -inf, -inf, -inf, -inf, -inf, -inf, -inf, -inf],[0., 0., 0., 0., 0., 0., -inf, -inf, -inf, -inf, -inf, -inf, -inf, -inf, -inf, -inf, -inf],[0., 0., 0., 0., 0., 0., 0., -inf, -inf, -inf, -inf, -inf, -inf, -inf, -inf, -inf, -inf],[0., 0., 0., 0., 0., 0., 0., 0., -inf, -inf, -inf, -inf, -inf, -inf, -inf, -inf, -inf],[0., 0., 0., 0., 0., 0., 0., 0., 0., -inf, -inf, -inf, -inf, -inf, -inf, -inf, -inf],[0., 0., 0., 0., 0., 0., 0., 0., 0., 0., -inf, -inf, -inf, -inf, -inf, -inf, -inf],[0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., -inf, -inf, -inf, -inf, -inf, -inf],[0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., -inf, -inf, -inf, -inf, -inf],[0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., -inf, -inf, -inf, -inf],[0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., -inf, -inf, -inf],[0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., -inf, -inf],[0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., -inf],[0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.]])
qk_per_token_after_masking = qk_per_token + maskdisplay_qk_heatmap (qk_per_token_after_masking)
qk_per_token_after_masking_after_softmax = torch.nn.functional.softmax (qk_per_token_after_masking, dim=1).to (torch.bfloat16) display_qk_heatmap (qk_per_token_after_masking_after_softmax)就像键一样，值权重也在 4 个注意力头之间共享（以节省计算量） 结果，下面的值权重矩阵形状为 [8x128x4096]
v_layer0 = model ["layers.0.attention.wv.weight"] v_layer0 = v_layer0.view (n_kv_heads, v_layer0.shape [0] //n_kv_heads, dim) v_layer0.shapetorch.Size ([8, 128, 4096])v_layer0_head0 = v_layer0 [0] v_layer0_head0.shapetorch.Size ([128, 4096])v_per_token = torch.matmul (token_embeddings, v_layer0_head0.T)v_per_token.shapetorch.Size ([17, 128])qkv_attention = torch.matmul (qk_per_token_after_masking_after_softmax, v_per_token) qkv_attention.shapetorch.Size ([17, 128])qkv_attention_store = []for head in range (n_heads):q_layer0_head = q_layer0 [head]k_layer0_head = k_layer0 [head//4] # key weights are shared across 4 headsv_layer0_head = v_layer0 [head//4] # value weights are shared across 4 headsq_per_token = torch.matmul (token_embeddings, q_layer0_head.T)k_per_token = torch.matmul (token_embeddings, k_layer0_head.T)v_per_token = torch.matmul (token_embeddings, v_layer0_head.T)q_per_token_split_into_pairs = q_per_token.float ().view (q_per_token.shape [0], -1, 2)q_per_token_as_complex_numbers = torch.view_as_complex (q_per_token_split_into_pairs)q_per_token_split_into_pairs_rotated = torch.view_as_real (q_per_token_as_complex_numbers * freqs_cis [:len (tokens)])q_per_token_rotated = q_per_token_split_into_pairs_rotated.view (q_per_token.shape)k_per_token_split_into_pairs = k_per_token.float ().view (k_per_token.shape [0], -1, 2)k_per_token_as_complex_numbers = torch.view_as_complex (k_per_token_split_into_pairs)k_per_token_split_into_pairs_rotated = torch.view_as_real (k_per_token_as_complex_numbers * freqs_cis [:len (tokens)])k_per_token_rotated = k_per_token_split_into_pairs_rotated.view (k_per_token.shape)qk_per_token = torch.matmul (q_per_token_rotated, k_per_token_rotated.T)/(128)**0.5mask = torch.full ((len (tokens), len (tokens)), float ("-inf"), device=tokens.device)mask = torch.triu (mask, diagonal=1)qk_per_token_after_masking = qk_per_token + maskqk_per_token_after_masking_after_softmax = torch.nn.functional.softmax (qk_per_token_after_masking, dim=1).to (torch.bfloat16)qkv_attention = torch.matmul (qk_per_token_after_masking_after_softmax, v_per_token)qkv_attention = torch.matmul (qk_per_token_after_masking_after_softmax, v_per_token)(qkv_attention)len (qkv_attention_store)
32stacked_qkv_attention = torch.cat (qkv_attention_store, dim=-1) stacked_qkv_attention.shapetorch.Size ([17, 4096])w_layer0 = model ["layers.0.attention.wo.weight"] w_layer0.shapetorch.Size ([4096, 4096])embedding_delta = torch.matmul (stacked_qkv_attention, w_layer0.T) embedding_delta.shapetorch.Size ([17, 4096])embedding_after_edit = token_embeddings_unnormalized + embedding_deltaembedding_after_edit.shape
torch.Size ([17, 4096])embedding_after_edit_normalized = rms_norm (embedding_after_edit, model ["layers.0.ffn_norm.weight"]) embedding_after_edit_normalized.shapetorch.Size ([17, 4096])w1 = model ["layers.0.feed_forward.w1.weight"] w2 = model ["layers.0.feed_forward.w2.weight"] w3 = model ["layers.0.feed_forward.w3.weight"] output_after_feedforward = torch.matmul (torch.functional.F.silu (torch.matmul (embedding_after_edit_normalized, w1.T)) * torch.matmul (embedding_after_edit_normalized, w3.T), w2.T) output_after_feedforward.shapetorch.Size ([17, 4096])layer_0_embedding = embedding_after_edit+output_after_feedforwardlayer_0_embedding.shape
torch.Size ([17, 4096])final_embedding = token_embeddings_unnormalizedfor layer in range (n_layers):qkv_attention_store = []layer_embedding_norm = rms_norm (final_embedding, model [f"layers.{layer}.attention_norm.weight"])q_layer = model [f"layers.{layer}.attention.wq.weight"]q_layer = q_layer.view (n_heads, q_layer.shape [0] //n_heads, dim)k_layer = model [f"layers.{layer}.attention.wk.weight"]k_layer = k_layer.view (n_kv_heads, k_layer.shape [0] //n_kv_heads, dim)v_layer = model [f"layers.{layer}.attention.wv.weight"]v_layer = v_layer.view (n_kv_heads, v_layer.shape [0] //n_kv_heads, dim)w_layer = model [f"layers.{layer}.attention.wo.weight"]for head in range (n_heads):q_layer_head = q_layer [head]k_layer_head = k_layer [head//4]v_layer_head = v_layer [head//4]q_per_token = torch.matmul (layer_embedding_norm, q_layer_head.T)k_per_token = torch.matmul (layer_embedding_norm, k_layer_head.T)v_per_token = torch.matmul (layer_embedding_norm, v_layer_head.T)q_per_token_split_into_pairs = q_per_token.float ().view (q_per_token.shape [0], -1, 2)q_per_token_as_complex_numbers = torch.view_as_complex (q_per_token_split_into_pairs)q_per_token_split_into_pairs_rotated = torch.view_as_real (q_per_token_as_complex_numbers * freqs_cis)q_per_token_rotated = q_per_token_split_into_pairs_rotated.view (q_per_token.shape)k_per_token_split_into_pairs = k_per_token.float ().view (k_per_token.shape [0], -1, 2)k_per_token_as_complex_numbers = torch.view_as_complex (k_per_token_split_into_pairs)k_per_token_split_into_pairs_rotated = torch.view_as_real (k_per_token_as_complex_numbers * freqs_cis)k_per_token_rotated = k_per_token_split_into_pairs_rotated.view (k_per_token.shape)qk_per_token = torch.matmul (q_per_token_rotated, k_per_token_rotated.T)/(128)**0.5mask = torch.full ((len (token_embeddings_unnormalized), len (token_embeddings_unnormalized)), float ("-inf"))mask = torch.triu (mask, diagonal=1)qk_per_token_after_masking = qk_per_token + maskqk_per_token_after_masking_after_softmax = torch.nn.functional.softmax (qk_per_token_after_masking, dim=1).to (torch.bfloat16)qkv_attention = torch.matmul (qk_per_token_after_masking_after_softmax, v_per_token)(qkv_attention)stacked_qkv_attention = torch.cat (qkv_attention_store, dim=-1)w_layer = model [f"layers.{layer}.attention.wo.weight"]embedding_delta = torch.matmul (stacked_qkv_attention, w_layer.T)embedding_after_edit = final_embedding + embedding_deltaembedding_after_edit_normalized = rms_norm (embedding_after_edit, model [f"layers.{layer}.ffn_norm.weight"])w1 = model [f"layers.{layer}.feed_forward.w1.weight"]w2 = model [f"layers.{layer}.feed_forward.w2.weight"]w3 = model [f"layers.{layer}.feed_forward.w3.weight"]output_after_feedforward = torch.matmul (torch.functional.F.silu (torch.matmul (embedding_after_edit_normalized, w1.T)) * torch.matmul (embedding_after_edit_normalized, w3.T), w2.T)final_embedding = embedding_after_edit+output_after_feedforward
final_embedding = rms_norm (final_embedding, model ["norm.weight"]) final_embedding.shapetorch.Size ([17, 4096])model ["output.weight"].shapetorch.Size ([128256, 4096])logits = torch.matmul (final_embedding [-1], model ["output.weight"].T) logits.shapetorch.Size ([128256])next_token = torch.argmax (logits, dim=-1) next_tokentensor (2983)tokenizer.decode ([next_token.item ()])'42'© THE END
转载请联系本公众号获得授权
投稿或寻求报道：[email protected]
微信扫码关注该文公众号作者
来源：机器之心
相关新闻
Karpathy新教程爆火，网友抢着送他H100：从头复现GPT-2训练智谱版Sora开源爆火：狂揽4K Star，4090单卡运行，A6000可微调Karpathy最新四小时视频教程：从零复现GPT-2，通宵运行即搞定爆火ChatTTS突破开源语音天花板，3天斩获9k的Star量1.5k star，搭建一个属于自己或团队的WIKI知识管理系统！19.2k star，三分钟实现一套 Web 版的 Excel 组件，这个开源工具强的离谱！大厂发函，3 万Star的开源项目被清空「后训练 + STaR」是 OpenAI 绝密项目「草莓」背后的秘密吗？14.2k star，一款开源的爬虫项目，支持小红书、抖音、微博等主流内容平台30K Star，最全面的PDF处理开源项目，你也可以拥有一个本地的PDF处理大全ControlNet作者又出爆款！一张图生成绘画全过程，两天狂揽1.4k Star大神Karpathy：我给大模型「SQL注入」攻击，简直不要太轻松Apple Intelligence存重大安全缺陷，几行代码即可攻破！Karpathy发文提醒ChatGPT「秘方」竟在拖LLM后腿？Karpathy、LeCun联手开怼RLHF！Karpathy观点惹争议：RLHF不是真正的强化学习，谷歌、Meta下场反对LLM智能「参差不齐」！AI大牛Karpathy用表情包解释「9.9<9.11」小模型狂飙！6家巨头争相发布小模型，Andrej Karpathy：大语言模型的尺寸竞争正在倒退...为什么AI数不清Strawberry里有几个 r？Karpathy：我用表情包给你解释一下贾扬清点赞：3K star量的SGLang上新，加速Llama 405B推理秒杀vLLM、TensorRT-LLMAI大神Karpathy官宣创业，打造80亿人AI原生学校！首款课程GitHub斩获17k星AI+教育！前OpenAI联创Andrej Karpathy官宣创业！创办第一所AI原生学校Karpathy又整活儿了！一天训练出GPT-2、成本还骤降100倍，网友：dream老黄把价格再打下来GitHub 8k Star，一作实习生，字节这个大模型成果被苹果选中了【allston】步行至BU, Allston天花板超大户型的高级公寓｜88超市和Star Market后｜出门就是绿线和公交