candidate changes to apis, have to think through more

README.md

@@ -42,6 +42,16 @@ $ python sample.py
 
 Training on 1 A100 40GB GPU overnight currently gets loss ~3.74, training on 4 gets ~3.60. Random chance at init is -ln(1/50257) = 10.82. Which brings us to baselines:
 
+## finetuning
+
+For an example of how to finetune a GPT on new text go to `data/shakespeare` and look at `prepare.py` to download the tiny shakespeare dataset and render it into a `train.bin` and `val.bin`. Unlike OpenWebText this will run in seconds. Finetuning takes very little time, e.g. on a single GPT just a few minutes. Run an example finetuning like:
+
+```
+$ python train.py finetune_shakespeare
+```
+
+This will load the config parameter overrides in `config/finetune_shakespeare.py` (I didn't tune them much though). Basically, we initialize from a GPT2 checkpoint with `init_from` and train as normal, except shorter and with a small learning rate. The best checkpoint (lowest validation loss) will be in the `out_dir` directory, e.g. in `out-shakespeare` by default, per the config file. You can then run the code in `sample.py` to generate infinite Shakespeare. Note that you'll have to edit it to point to the correct `out_dir`.
+
 ## baselines
 
 OpenAI GPT-2 checkpoints allow us to get some baselines in place for openwebtext. We can get the numbers as follows:

config/finetune_shakespeare.py

@@ -0,0 +1,22 @@
+import time
+
+out_dir = 'out-shakespeare'
+eval_interval = 200
+wandb_log = False # feel free to turn on
+wandb_project = 'shakespeare'
+wandb_run_name = 'ft-' + str(time.time())
+compile_model = False # takes too little time to finetune, not worth it
+
+# save a nice and overfit checkpoint that
+# will only speak Shakespeare and forgets
+# everything else about the world #dark
+always_save_checkpoint = True
+
+dataset = 'shakespeare'
+init_from = 'gpt2-xl'
+batch_size = 1
+block_size = 512
+
+learning_rate = 1e-5
+max_iters = 1000
+decay_lr = False

data/openwebtext/prepare.py

@@ -35,6 +35,7 @@ enc = tiktoken.get_encoding("gpt2")
 def process(example):
     ids = enc.encode_ordinary(example['text']) # encode_ordinary ignores any special tokens
     ids.append(enc.eot_token) # add the end of text token, e.g. 50256 for gpt2 bpe
+    # note: I think eot should be prepended not appended... hmm. it's called "eot" though...
     out = {'ids': ids, 'len': len(ids)}
     return out
 

data/shakespeare/prepare.py

@@ -0,0 +1,32 @@
+import os
+import requests
+import tiktoken
+import numpy as np
+
+# download the tiny shakespeare dataset
+if not os.path.exists('input.txt'):
+    data_url = 'https://raw.githubusercontent.com/karpathy/char-rnn/master/data/tinyshakespeare/input.txt'
+    with open('input.txt', 'w') as f:
+        f.write(requests.get(data_url).text)
+
+with open('input.txt', 'r') as f:
+    data = f.read()
+n = len(data)
+train_data = data[:int(n*0.9)]
+val_data = data[int(n*0.9):]
+
+# encode with tiktoken gpt2 bpe
+enc = tiktoken.get_encoding("gpt2")
+train_ids = enc.encode_ordinary(train_data)
+val_ids = enc.encode_ordinary(val_data)
+print(f"train has {len(train_ids)} tokens")
+print(f"val has {len(val_ids)} tokens")
+
+# export to bin files
+train_ids = np.array(train_ids, dtype=np.uint16)
+val_ids = np.array(val_ids, dtype=np.uint16)
+train_ids.tofile('train.bin')
+val_ids.tofile('val.bin')
+
+# train.bin has 301,966 tokens
+# val.bin has 36,059 tokens

data/shakespeare/readme.md

@@ -0,0 +1,9 @@
+
+# tiny shakespeare
+
+Tiny shakespeare, of the good old char-rnn fame :)
+
+After running `prepare.py`:
+
+- train.bin has 301,966 tokens
+- val.bin has 36,059 tokens

model.py

@@ -90,7 +90,7 @@ class Block(nn.Module):
         x = x + self.mlp(self.ln_2(x))
         return x
 
-@dataclass(frozen=True)
+@dataclass
 class GPTConfig:
     block_size: int = 1024
     vocab_size: int = 50257
@@ -105,7 +105,7 @@ class GPT(nn.Module):
         super().__init__()
         assert config.vocab_size is not None
         assert config.block_size is not None
-        self.block_size = config.block_size
+        self.config = config
 
         self.transformer = nn.ModuleDict(dict(
             wte = nn.Embedding(config.vocab_size, config.n_embd),
@@ -123,7 +123,7 @@ class GPT(nn.Module):
     def forward(self, idx, targets=None):
         device = idx.device
         b, t = idx.size()
-        assert t <= self.block_size, f"Cannot forward sequence of length {t}, block size is only {self.block_size}"
+        assert t <= self.config.block_size, f"Cannot forward sequence of length {t}, block size is only {self.config.block_size}"
         pos = torch.arange(0, t, dtype=torch.long, device=device).unsqueeze(0) # shape (1, t)
 
         # forward the GPT model itself
@@ -146,27 +146,36 @@ class GPT(nn.Module):
         # model surgery to decrease the block size if necessary
         # e.g. we may load the GPT2 pretrained model checkpoint (block size 1024)
         # but want to use a smaller block size for some smaller, simpler model
-        assert block_size <= self.block_size
-        self.block_size = block_size
+        assert block_size <= self.config.block_size
+        self.config.block_size = block_size
         self.transformer.wpe.weight = nn.Parameter(self.transformer.wpe.weight[:block_size])
         for block in self.transformer.h:
             block.attn.bias = block.attn.bias[:,:,:block_size,:block_size]
 
     @classmethod
-    def from_pretrained(cls, model_type):
+    def from_pretrained(cls, model_type, override_args):
         assert model_type in {'gpt2', 'gpt2-medium', 'gpt2-large', 'gpt2-xl'}
+        # only dropout can be overridden see more notes below
+        assert all(k == 'dropout' for k in override_args)
         from transformers import GPT2LMHeadModel
         print("loading weights from pretrained gpt: %s" % model_type)
 
-        layer_config = {
+        # n_layer, n_head and n_embd are determined from model_type
+        config_args = {
             'gpt2':         dict(n_layer=12, n_head=12, n_embd=768),  # 124M params
             'gpt2-medium':  dict(n_layer=24, n_head=16, n_embd=1024), # 350M params
             'gpt2-large':   dict(n_layer=36, n_head=20, n_embd=1280), # 774M params
             'gpt2-xl':      dict(n_layer=48, n_head=25, n_embd=1600), # 1558M params
         }[model_type]
+        # we can override the dropout rate
+        if 'dropout' in override_args:
+            config_args['dropout'] = override_args['dropout']
+        # block_size is always 1024 for GPT model checkpoints
+        # if one wants a lower block_size it has to be done through model surgery
+        # later, by calling crop_block_size()
 
         # create a from-scratch initialized minGPT model
-        config = GPTConfig(block_size=1024, **layer_config)
+        config = GPTConfig(block_size=1024, **config_args)
         model = GPT(config)
         sd = model.state_dict()
 
@@ -248,7 +257,7 @@ class GPT(nn.Module):
         """
         for _ in range(max_new_tokens):
             # if the sequence context is growing too long we must crop it at block_size
-            idx_cond = idx if idx.size(1) <= self.block_size else idx[:, -self.block_size:]
+            idx_cond = idx if idx.size(1) <= self.config.block_size else idx[:, -self.config.block_size:]
             # forward the model to get the logits for the index in the sequence
             logits, _ = self(idx_cond)
             # pluck the logits at the final step and scale by desired temperature

sample.py

@@ -21,18 +21,18 @@ model = GPT(gptconf)
 model.load_state_dict(checkpoint['model'])
 model.eval()
 model.to(device)
-model = torch.compile(model) # requires PyTorch 2.0
+#model = torch.compile(model) # requires PyTorch 2.0 (optional)
 
 enc = tiktoken.get_encoding("gpt2")
-#start = enc.encode("\n")
-start = [enc.eot_token]
+start = enc.encode("\n") # user choice on what token to start with
+#start = [enc.eot_token]
 x = (torch.tensor(start, dtype=torch.long, device=device)[None, ...])
 
-for k in range(1):
+for k in range(10):
 
     with torch.no_grad():
         with torch.amp.autocast(device_type="cuda", dtype=torch.bfloat16):
-            y = model.generate(x, 300, temperature=0.8, top_k=200)
+            y = model.generate(x, 500, temperature=0.8, top_k=200)
 
     print(enc.decode(y[0].tolist()))
     print('---------------')

train.py

@@ -31,6 +31,7 @@ eval_interval = 500
 log_interval = 1
 eval_iters = 50
 eval_only = False # if True, script exits right after the first eval
+always_save_checkpoint = False # if True, always save a checkpoint after each eval
 # wandb logging
 wandb_log = False # disabled by default
 wandb_entity = 'karpathy'
@@ -138,6 +139,7 @@ elif init_from == 'resume':
     checkpoint_model_args = checkpoint['model_args']
     for k, v in model_args.items():
         assert checkpoint_model_args[k] == v, "for now"
+        # TODO: think through how passed in params should interact with checkpoint params
     gptconf = GPTConfig(**model_args)
     model = GPT(gptconf)
     model.load_state_dict(checkpoint['model'])
@@ -146,9 +148,14 @@ elif init_from == 'resume':
 elif init_from.startswith('gpt2'):
     print(f"Initializing from OpenAI GPT-2 weights: {init_from}")
     # initialize from OpenAI GPT-2 weights
-    model = GPT.from_pretrained(init_from)
+    override_args = dict(dropout=dropout)
+    model = GPT.from_pretrained(init_from, override_args)
+    # read off and override the GPT sizing model args from the model config
+    model_args['n_layer'] = model.config.n_layer
+    model_args['n_head'] = model.config.n_head
+    model_args['n_embd'] = model.config.n_embd
 # crop down the model block size if desired
-if block_size < model.block_size:
+if block_size < model.config.block_size:
     model.crop_block_size(block_size)
 model.to(device)
 
@@ -227,7 +234,7 @@ while True:
                 "val/loss": losses['val'],
                 "lr": lr,
             })
-        if losses['val'] < best_val_loss:
+        if losses['val'] < best_val_loss or always_save_checkpoint:
             best_val_loss = losses['val']
             raw_model = model.module if ddp else model
             if iter_num > 0:
@@ -238,6 +245,7 @@ while True:
                     'iter_num': iter_num,
                     'best_val_loss': best_val_loss,
                 }
+                print(f"saving checkpoint to {out_dir}")
                 torch.save(checkpoint, os.path.join(out_dir, 'ckpt.pt'))
     if iter_num == 0 and eval_only:
         break
@@ -260,7 +268,8 @@ while True:
     iter_num += 1
 
     # termination conditions
-    if iter_num >= max_iters:
+    if iter_num > max_iters:
         break
 
-destroy_process_group()
+if ddp:
+    destroy_process_group()