Add general name to train

[teleprint-me]

This commit adds the model name to a GGML trained model when using train-text-from-scratch.

19:41:23 | /mnt/valerie/forked/ggerganov/llama.cpp
 git:(add-general-name-to-train | θ) λ python gguf-py/scripts/gguf-dump.py models/valerie/v0.1/ggml-valerie-v0.1-256x32-f32-LATEST.gguf --no-tensors
* Loading: models/valerie/v0.1/ggml-valerie-v0.1-256x32-f32-LATEST.gguf
* File is LITTLE endian, script is running on a LITTLE endian host.

* Dumping 24 key/value pair(s)
      1: UINT32     |        1 | GGUF.version = 3
      2: UINT64     |        1 | GGUF.tensor_count = 147
      3: UINT64     |        1 | GGUF.kv_count = 21
      4: STRING     |        1 | general.architecture = 'llama'
      5: STRING     |        1 | general.name = 'llama'  # Adds the models name
      6: UINT32     |        1 | general.file_type = 0
      7: UINT32     |        1 | llama.context_length = 256
      8: UINT32     |        1 | llama.embedding_length = 256
      9: UINT32     |        1 | llama.feed_forward_length = 768
     10: UINT32     |        1 | llama.attention.head_count = 8
     11: UINT32     |        1 | llama.block_count = 16
     12: UINT32     |        1 | llama.rope.dimension_count = 32
     13: FLOAT32    |        1 | llama.attention.layer_norm_rms_epsilon = 9.999999747378752e-06
     14: FLOAT32    |        1 | llama.rope.freq_base = 10000.0
     15: FLOAT32    |        1 | llama.rope.scale_linear = 1.0
     16: STRING     |        1 | tokenizer.ggml.model = 'llama'
     17: [FLOAT32]  |    32000 | tokenizer.ggml.scores
     18: [INT32]    |    32000 | tokenizer.ggml.token_type
     19: [STRING]   |    32000 | tokenizer.ggml.tokens
     20: UINT32     |        1 | tokenizer.ggml.bos_token_id = 1
     21: UINT32     |        1 | tokenizer.ggml.eos_token_id = 2
     22: UINT32     |        1 | tokenizer.ggml.unknown_token_id = 0
     23: UINT32     |        1 | tokenizer.ggml.seperator_token_id = 4294967295
     24: UINT32     |        1 | tokenizer.ggml.padding_token_id = 4294967295

This commit simply uses the models architecture as a base to keep the changes both minimal and simple until I have time to come up with a more customizable approach.